{
  "runs": [
    {
      "id": "8ov14yyc",
      "url": "https://wandb.ai/hzj-lab/shopsimrl/runs/8ov14yyc",
      "round": 0,
      "primary": true,
      "state": "finished",
      "created_at": "2026-09-06T10:26:00Z",
      "history_rows": 100,
      "keys": [
        "_runtime",
        "_step",
        "_timestamp",
        "perf/actor_train_tflops",
        "perf/actor_train_time",
        "perf/actor_train_tok_per_s",
        "perf/data_preprocess_time",
        "perf/effective_tokens_per_gpu_per_sec",
        "perf/log_probs_tflops",
        "perf/log_probs_time",
        "perf/longest_effective_sample_tokens_per_sec",
        "perf/longest_sample_tokens_per_sec",
        "perf/rollout_time",
        "perf/sleep_time",
        "perf/step_time",
        "perf/tokens_per_gpu_per_sec",
        "perf/train_time",
        "perf/train_wait_time",
        "perf/update_weights_time",
        "perf/wait_time_ratio",
        "perf/wake_up_time",
        "rollout/advantages",
        "rollout/avg_cached_tokens_per_sample",
        "rollout/dynamic_filter/drop_shopsim_unscored_technical_group",
        "rollout/dynamic_filter/drop_shopsim_zero_std_group",
        "rollout/kl",
        "rollout/log_probs",
        "rollout/prefix_cache_hit_rate",
        "rollout/raw_reward",
        "rollout/repetition_frac",
        "rollout/response_len/max",
        "rollout/response_len/mean",
        "rollout/response_len/median",
        "rollout/response_len/min",
        "rollout/response_lengths",
        "rollout/returns",
        "rollout/rewards",
        "rollout/rollout_log_probs",
        "rollout/shopsim/all_wrong_group_rate",
        "rollout/shopsim/assisted_empty_group_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho",
        "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate",
        "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho",
        "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate",
        "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho",
        "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate",
        "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho",
        "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate",
        "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho",
        "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate",
        "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho",
        "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate",
        "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho",
        "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate",
        "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho",
        "rollout/shopsim/full_skill_retry_rate",
        "rollout/shopsim/full_skill_retry_success_rate",
        "rollout/shopsim/group_count",
        "rollout/shopsim/internalization_deficit_rate",
        "rollout/shopsim/pending_analysis_rate",
        "rollout/shopsim/q",
        "rollout/shopsim/reward_mean",
        "rollout/shopsim/rho_max",
        "rollout/shopsim/rho_mean",
        "rollout/shopsim/rho_min",
        "rollout/shopsim/runtime_error_group_rate",
        "rollout/shopsim/scored_rate",
        "rollout/shopsim/skill_free_group_rate",
        "rollout/shopsim/skills_per_assisted_group_mean",
        "rollout/shopsim/strict_reward_mean",
        "rollout/shopsim/success_rate",
        "rollout/shopsim/technical_group_drop_count",
        "rollout/shopsim/technical_group_drop_rate",
        "rollout/shopsim/train_group_count",
        "rollout/shopsim/train_trajectory_count",
        "rollout/shopsim/trajectory_count",
        "rollout/shopsim/zero_std_group_drop_count",
        "rollout/shopsim/zero_std_group_drop_rate",
        "rollout/step",
        "rollout/total_lengths",
        "rollout/truncated",
        "rollout/truncated_ratio",
        "rollout/zero_std/count_0.0",
        "rollout/zero_std/count_0.5",
        "rollout/zero_std/count_0.7",
        "rollout/zero_std/count_1.0",
        "train/entropy_loss",
        "train/global_batch_size",
        "train/grad_norm",
        "train/loss",
        "train/lr-pg_0",
        "train/lr-pg_1",
        "train/pg_clipfrac",
        "train/pg_loss",
        "train/ppo_kl",
        "train/step",
        "train/train_rollout_logprob_abs_diff"
      ],
      "sha256": "f38c31eac9be4fa4286680dcc96241d3f3936233351301f45da25420885eead6"
    },
    {
      "id": "chjizc8u",
      "url": "https://wandb.ai/hzj-lab/shopsimrl/runs/chjizc8u",
      "round": 1,
      "primary": true,
      "state": "finished",
      "created_at": "2026-09-06T15:04:16Z",
      "history_rows": 100,
      "keys": [
        "_runtime",
        "_step",
        "_timestamp",
        "perf/actor_train_tflops",
        "perf/actor_train_time",
        "perf/actor_train_tok_per_s",
        "perf/data_preprocess_time",
        "perf/effective_tokens_per_gpu_per_sec",
        "perf/log_probs_tflops",
        "perf/log_probs_time",
        "perf/longest_effective_sample_tokens_per_sec",
        "perf/longest_sample_tokens_per_sec",
        "perf/rollout_time",
        "perf/sleep_time",
        "perf/step_time",
        "perf/tokens_per_gpu_per_sec",
        "perf/train_time",
        "perf/train_wait_time",
        "perf/update_weights_time",
        "perf/wait_time_ratio",
        "perf/wake_up_time",
        "rollout/advantages",
        "rollout/avg_cached_tokens_per_sample",
        "rollout/dynamic_filter/drop_shopsim_zero_std_group",
        "rollout/kl",
        "rollout/log_probs",
        "rollout/prefix_cache_hit_rate",
        "rollout/raw_reward",
        "rollout/repetition_frac",
        "rollout/response_len/max",
        "rollout/response_len/mean",
        "rollout/response_len/median",
        "rollout/response_len/min",
        "rollout/response_lengths",
        "rollout/returns",
        "rollout/rewards",
        "rollout/rollout_log_probs",
        "rollout/shopsim/all_wrong_group_rate",
        "rollout/shopsim/assisted_empty_group_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho",
        "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate",
        "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho",
        "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate",
        "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho",
        "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate",
        "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho",
        "rollout/shopsim/full_skill_retry_rate",
        "rollout/shopsim/full_skill_retry_success_rate",
        "rollout/shopsim/group_count",
        "rollout/shopsim/internalization_deficit_rate",
        "rollout/shopsim/pending_analysis_rate",
        "rollout/shopsim/q",
        "rollout/shopsim/reward_mean",
        "rollout/shopsim/rho_max",
        "rollout/shopsim/rho_mean",
        "rollout/shopsim/rho_min",
        "rollout/shopsim/runtime_error_group_rate",
        "rollout/shopsim/scored_rate",
        "rollout/shopsim/skill_free_group_rate",
        "rollout/shopsim/skills_per_assisted_group_mean",
        "rollout/shopsim/strict_reward_mean",
        "rollout/shopsim/success_rate",
        "rollout/shopsim/train_group_count",
        "rollout/shopsim/train_trajectory_count",
        "rollout/shopsim/trajectory_count",
        "rollout/shopsim/zero_std_group_drop_count",
        "rollout/shopsim/zero_std_group_drop_rate",
        "rollout/step",
        "rollout/total_lengths",
        "rollout/truncated",
        "rollout/truncated_ratio",
        "rollout/zero_std/count_0.0",
        "rollout/zero_std/count_0.4",
        "rollout/zero_std/count_0.5",
        "rollout/zero_std/count_1.0",
        "train/entropy_loss",
        "train/global_batch_size",
        "train/grad_norm",
        "train/loss",
        "train/lr-pg_0",
        "train/lr-pg_1",
        "train/pg_clipfrac",
        "train/pg_loss",
        "train/ppo_kl",
        "train/step",
        "train/train_rollout_logprob_abs_diff"
      ],
      "sha256": "7052dc3a60e8d4925f737c879b8bd6deb0f6ad18f2a4aec1cd3760cd2adc2e86"
    },
    {
      "id": "7vhbv5j0",
      "url": "https://wandb.ai/hzj-lab/shopsimrl/runs/7vhbv5j0",
      "round": 2,
      "primary": true,
      "state": "finished",
      "created_at": "2026-09-07T01:49:09Z",
      "history_rows": 100,
      "keys": [
        "_runtime",
        "_step",
        "_timestamp",
        "perf/actor_train_tflops",
        "perf/actor_train_time",
        "perf/actor_train_tok_per_s",
        "perf/data_preprocess_time",
        "perf/effective_tokens_per_gpu_per_sec",
        "perf/log_probs_tflops",
        "perf/log_probs_time",
        "perf/longest_effective_sample_tokens_per_sec",
        "perf/longest_sample_tokens_per_sec",
        "perf/rollout_time",
        "perf/sleep_time",
        "perf/step_time",
        "perf/tokens_per_gpu_per_sec",
        "perf/train_time",
        "perf/train_wait_time",
        "perf/update_weights_time",
        "perf/wait_time_ratio",
        "perf/wake_up_time",
        "rollout/advantages",
        "rollout/avg_cached_tokens_per_sample",
        "rollout/dynamic_filter/drop_shopsim_zero_std_group",
        "rollout/kl",
        "rollout/log_probs",
        "rollout/prefix_cache_hit_rate",
        "rollout/raw_reward",
        "rollout/repetition_frac",
        "rollout/response_len/max",
        "rollout/response_len/mean",
        "rollout/response_len/median",
        "rollout/response_len/min",
        "rollout/response_lengths",
        "rollout/returns",
        "rollout/rewards",
        "rollout/rollout_log_probs",
        "rollout/shopsim/all_wrong_group_rate",
        "rollout/shopsim/assisted_empty_group_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho",
        "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate",
        "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho",
        "rollout/shopsim/full_skill_retry_rate",
        "rollout/shopsim/full_skill_retry_success_rate",
        "rollout/shopsim/group_count",
        "rollout/shopsim/internalization_deficit_rate",
        "rollout/shopsim/pending_analysis_rate",
        "rollout/shopsim/q",
        "rollout/shopsim/reward_mean",
        "rollout/shopsim/rho_max",
        "rollout/shopsim/rho_mean",
        "rollout/shopsim/rho_min",
        "rollout/shopsim/runtime_error_group_rate",
        "rollout/shopsim/scored_rate",
        "rollout/shopsim/skill_free_group_rate",
        "rollout/shopsim/skills_per_assisted_group_mean",
        "rollout/shopsim/strict_reward_mean",
        "rollout/shopsim/success_rate",
        "rollout/shopsim/train_group_count",
        "rollout/shopsim/train_trajectory_count",
        "rollout/shopsim/trajectory_count",
        "rollout/shopsim/zero_std_group_drop_count",
        "rollout/shopsim/zero_std_group_drop_rate",
        "rollout/step",
        "rollout/total_lengths",
        "rollout/truncated",
        "rollout/truncated_ratio",
        "rollout/zero_std/count_0.0",
        "rollout/zero_std/count_0.5",
        "rollout/zero_std/count_1.0",
        "train/entropy_loss",
        "train/global_batch_size",
        "train/grad_norm",
        "train/loss",
        "train/lr-pg_0",
        "train/lr-pg_1",
        "train/pg_clipfrac",
        "train/pg_loss",
        "train/ppo_kl",
        "train/step",
        "train/train_rollout_logprob_abs_diff"
      ],
      "sha256": "4012c89e093afc5ecce65a340946a4ff48574a3de71714ac1ddad0ccd300892c"
    },
    {
      "id": "y4xk6byf",
      "url": "https://wandb.ai/hzj-lab/shopsimrl/runs/y4xk6byf",
      "round": 3,
      "primary": true,
      "state": "finished",
      "created_at": "2026-09-07T07:15:56Z",
      "history_rows": 100,
      "keys": [
        "_runtime",
        "_step",
        "_timestamp",
        "perf/actor_train_tflops",
        "perf/actor_train_time",
        "perf/actor_train_tok_per_s",
        "perf/data_preprocess_time",
        "perf/effective_tokens_per_gpu_per_sec",
        "perf/log_probs_tflops",
        "perf/log_probs_time",
        "perf/longest_effective_sample_tokens_per_sec",
        "perf/longest_sample_tokens_per_sec",
        "perf/rollout_time",
        "perf/sleep_time",
        "perf/step_time",
        "perf/tokens_per_gpu_per_sec",
        "perf/train_time",
        "perf/train_wait_time",
        "perf/update_weights_time",
        "perf/wait_time_ratio",
        "perf/wake_up_time",
        "rollout/advantages",
        "rollout/avg_cached_tokens_per_sample",
        "rollout/dynamic_filter/drop_shopsim_zero_std_group",
        "rollout/kl",
        "rollout/log_probs",
        "rollout/prefix_cache_hit_rate",
        "rollout/raw_reward",
        "rollout/repetition_frac",
        "rollout/response_len/max",
        "rollout/response_len/mean",
        "rollout/response_len/median",
        "rollout/response_len/min",
        "rollout/response_lengths",
        "rollout/returns",
        "rollout/rewards",
        "rollout/rollout_log_probs",
        "rollout/shopsim/all_wrong_group_rate",
        "rollout/shopsim/assisted_empty_group_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate",
        "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate",
        "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate",
        "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho",
        "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate",
        "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho",
        "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate",
        "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho",
        "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate",
        "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate",
        "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho",
        "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate",
        "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho",
        "rollout/shopsim/full_skill_retry_rate",
        "rollout/shopsim/full_skill_retry_success_rate",
        "rollout/shopsim/group_count",
        "rollout/shopsim/internalization_deficit_rate",
        "rollout/shopsim/pending_analysis_rate",
        "rollout/shopsim/q",
        "rollout/shopsim/reward_mean",
        "rollout/shopsim/rho_max",
        "rollout/shopsim/rho_mean",
        "rollout/shopsim/rho_min",
        "rollout/shopsim/runtime_error_group_rate",
        "rollout/shopsim/scored_rate",
        "rollout/shopsim/skill_free_group_rate",
        "rollout/shopsim/skills_per_assisted_group_mean",
        "rollout/shopsim/strict_reward_mean",
        "rollout/shopsim/success_rate",
        "rollout/shopsim/train_group_count",
        "rollout/shopsim/train_trajectory_count",
        "rollout/shopsim/trajectory_count",
        "rollout/shopsim/zero_std_group_drop_count",
        "rollout/shopsim/zero_std_group_drop_rate",
        "rollout/step",
        "rollout/total_lengths",
        "rollout/truncated",
        "rollout/truncated_ratio",
        "rollout/zero_std/count_0.0",
        "rollout/zero_std/count_0.5",
        "rollout/zero_std/count_1.0",
        "train/entropy_loss",
        "train/global_batch_size",
        "train/grad_norm",
        "train/loss",
        "train/lr-pg_0",
        "train/lr-pg_1",
        "train/pg_clipfrac",
        "train/pg_loss",
        "train/ppo_kl",
        "train/step",
        "train/train_rollout_logprob_abs_diff"
      ],
      "sha256": "f48ecf5e2ba5fc0a77f2b845c60c8a8d9178237c21ff3dd742cbc18762309a56"
    }
  ],
  "steps": [
    {
      "step": 1,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 253,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 12,
      "rollout/shopsim/strict_reward_mean": 0.5843454790823213,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 0.96484375,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.46153846153846156,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.42857142857142855,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.38461538461538464,
      "rollout/truncated_ratio": 0.17654808959156784,
      "rollout/step": 1,
      "rollout/shopsim/reward_mean": 0.5843454790823213,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 8.016749281502413,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 47,
      "rollout/response_len/mean": 337.3017127799736,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.11538461538461539,
      "perf/rollout_time": 288.39619636535645,
      "rollout/shopsim/success_rate": 0.562753036437247,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/runtime_error_group_rate": 0.125,
      "rollout/shopsim/all_wrong_group_rate": 0.10714285714285714,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 8.016749281502413,
      "rollout/shopsim/zero_std_group_drop_count": 12,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.538461538461538,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.8076923076923077,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.3076923076923077,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9615384615384616,
      "perf/tokens_per_gpu_per_sec": 221.9273374844286,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 2312,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/technical_group_drop_count": 4,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/dynamic_filter/drop_shopsim_unscored_technical_group": 4,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/technical_group_drop_rate": 0.2,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5384615384615384,
      "perf/effective_tokens_per_gpu_per_sec": 221.9273374844286,
      "rollout/log_probs": -0.19044086337089539,
      "rollout/returns": -5.587935447692871e-09,
      "rollout/raw_reward": 0.35111989459815546,
      "rollout/kl": 0,
      "rollout/rewards": -0.12082328611054753,
      "rollout/rollout_log_probs": -0.18988516926765442,
      "rollout/total_lengths": 7767.442687747036,
      "rollout/advantages": -5.587935447692871e-09,
      "rollout/response_lengths": 337.3017127799736,
      "rollout/truncated": 0.17654808959156784,
      "perf/actor_train_tok_per_s": 13393.775002992987,
      "perf/sleep_time": 3.088714838027954,
      "perf/data_preprocess_time": 0.11687445640563965,
      "perf/wait_time_ratio": 0.34331557812307995,
      "perf/log_probs_time": 120.5582013130188,
      "perf/train_time": 561.309506893158,
      "perf/wake_up_time": 2.1805977821350098,
      "perf/train_wait_time": 293.4534327983856,
      "perf/actor_train_time": 440.16634583473206,
      "perf/step_time": 854.7629396915436,
      "perf/actor_train_tflops": 102.01074000109944,
      "perf/log_probs_tflops": 124.14942650976671,
      "perf/update_weights_time": 1.7349004745483398
    },
    {
      "step": 2,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 239.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.5230034722222222,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.20833333333333334,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.25,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5833333333333334,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.3333333333333333,
      "rollout/truncated_ratio": 0.3668032786885246,
      "rollout/step": 2,
      "rollout/shopsim/reward_mean": 0.5230034722222222,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 14.062823626671815,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 23,
      "rollout/response_len/mean": 308.0686475409836,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.16666666666666666,
      "perf/rollout_time": 291.26440811157227,
      "rollout/shopsim/success_rate": 0.515625,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.1875,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.25,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 14.062823626671815,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0.25,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.791666666666667,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.4166666666666667,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5416666666666666,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9166666666666666,
      "perf/tokens_per_gpu_per_sec": 258.07736169125656,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0.0625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.4583333333333333,
      "perf/effective_tokens_per_gpu_per_sec": 258.07736169125656,
      "rollout/log_probs": -0.20873617380857468,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/raw_reward": 0.25967668488160295,
      "rollout/kl": 0,
      "rollout/rewards": -0.21369923541291816,
      "rollout/rollout_log_probs": -0.20742356777191162,
      "rollout/total_lengths": 9558.578893442624,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/response_lengths": 308.0686475409836,
      "rollout/truncated": 0.3668032786885246,
      "perf/actor_train_tok_per_s": 13251.109144790933,
      "perf/sleep_time": 13.372366905212402,
      "perf/data_preprocess_time": 0.14818835258483887,
      "perf/wait_time_ratio": 0.2617174011747809,
      "perf/log_probs_time": 185.3799226284027,
      "perf/train_time": 890.229240655899,
      "perf/wake_up_time": 7.296461343765259,
      "perf/train_wait_time": 315.58170771598816,
      "perf/actor_train_time": 704.0295946598053,
      "perf/step_time": 1205.8109483718872,
      "perf/actor_train_tflops": 106.78051195971805,
      "perf/log_probs_tflops": 135.175804525606,
      "perf/update_weights_time": 1.7194387912750244
    },
    {
      "step": 3,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 276.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.5738932291666666,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.25925925925925924,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/full_skill_retry_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.6296296296296297,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.25925925925925924,
      "rollout/truncated_ratio": 0.21031746031746032,
      "rollout/step": 3,
      "rollout/shopsim/reward_mean": 0.5738932291666666,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 14.698668590364575,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 33,
      "rollout/response_len/mean": 341.8862433862434,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.2962962962962963,
      "perf/rollout_time": 278.66469502449036,
      "rollout/shopsim/success_rate": 0.5625,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 14.698668590364575,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0.16666666666666666,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.888888888888889,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5185185185185185,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 1,
      "perf/tokens_per_gpu_per_sec": 231.87903295148746,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0.03125,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.48148148148148145,
      "perf/effective_tokens_per_gpu_per_sec": 231.87903295148746,
      "rollout/log_probs": -0.20363140851259232,
      "rollout/returns": 3.5390257835388184e-08,
      "rollout/raw_reward": 0.42829585537918874,
      "rollout/kl": 0,
      "rollout/rewards": -0.1422389942817587,
      "rollout/rollout_log_probs": -0.20278357714414597,
      "rollout/total_lengths": 8056.920634920635,
      "rollout/advantages": 3.5390257835388184e-08,
      "rollout/response_lengths": 341.8862433862434,
      "rollout/truncated": 0.21031746031746032,
      "perf/actor_train_tok_per_s": 13412.286567453591,
      "perf/sleep_time": 12.906272411346436,
      "perf/data_preprocess_time": 0.11009931564331055,
      "perf/wait_time_ratio": 0.34463394137818965,
      "perf/log_probs_time": 118.868492603302,
      "perf/train_time": 574.347496509552,
      "perf/wake_up_time": 7.08655309677124,
      "perf/train_wait_time": 302.0291314125061,
      "perf/actor_train_time": 454.13822388648987,
      "perf/step_time": 876.3766279220581,
      "perf/actor_train_tflops": 103.41141896171864,
      "perf/log_probs_tflops": 131.69477492404354,
      "perf/update_weights_time": 1.736161470413208
    },
    {
      "step": 4,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 242,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.5943080357142857,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.16666666666666666,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5833333333333334,
      "rollout/shopsim/full_skill_retry_rate": 0.28125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.7083333333333334,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.16666666666666666,
      "rollout/truncated_ratio": 0.16290726817042606,
      "rollout/step": 4,
      "rollout/shopsim/reward_mean": 0.5943080357142857,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 8.419898733959688,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 50,
      "rollout/response_len/mean": 320.85588972431077,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.16666666666666666,
      "perf/rollout_time": 259.3855423927307,
      "rollout/shopsim/success_rate": 0.58984375,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.21875,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.28125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 8.419898733959688,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0.2222222222222222,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.333333333333333,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.625,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.3333333333333333,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.375,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.8333333333333334,
      "perf/tokens_per_gpu_per_sec": 246.77840333553573,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 2184,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/internalization_deficit_rate": 0.0625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.375,
      "perf/effective_tokens_per_gpu_per_sec": 246.77840333553573,
      "rollout/log_probs": -0.1887982040643692,
      "rollout/returns": 1.4901161193847656e-08,
      "rollout/raw_reward": 0.33154314357321873,
      "rollout/kl": 0,
      "rollout/rewards": -0.15889760089996166,
      "rollout/rollout_log_probs": -0.18809824436903,
      "rollout/total_lengths": 8063.015037593985,
      "rollout/advantages": 1.4901161193847656e-08,
      "rollout/response_lengths": 320.85588972431077,
      "rollout/truncated": 0.16290726817042606,
      "perf/actor_train_tok_per_s": 13448.728609649675,
      "perf/sleep_time": 12.474005937576294,
      "perf/data_preprocess_time": 0.11857819557189941,
      "perf/wait_time_ratio": 0.3171860337593439,
      "perf/log_probs_time": 127.01587128639221,
      "perf/train_time": 607.8104777336121,
      "perf/wake_up_time": 7.207159519195557,
      "perf/train_wait_time": 282.34483218193054,
      "perf/actor_train_time": 478.4308009147644,
      "perf/step_time": 890.1553099155426,
      "perf/actor_train_tflops": 103.8234029673303,
      "perf/log_probs_tflops": 130.35723628416028,
      "perf/update_weights_time": 1.7250373363494873
    },
    {
      "step": 5,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 228.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.4833984375,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.18518518518518517,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.37037037037037035,
      "rollout/shopsim/full_skill_retry_rate": 0.3125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.4444444444444444,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.2962962962962963,
      "rollout/truncated_ratio": 0.19444444444444445,
      "rollout/step": 5,
      "rollout/shopsim/reward_mean": 0.4833984375,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 13.208522605932743,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 43,
      "rollout/response_len/mean": 311.92270531400965,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.07407407407407407,
      "perf/rollout_time": 310.1028118133545,
      "rollout/shopsim/success_rate": 0.46875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.25,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.3125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 13.208522605932743,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0.2,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.518518518518518,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5925925925925926,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5185185185185185,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9629629629629629,
      "perf/tokens_per_gpu_per_sec": 208.2148163134437,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0.0625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5925925925925926,
      "perf/effective_tokens_per_gpu_per_sec": 208.2148163134437,
      "rollout/log_probs": -0.17881229519844055,
      "rollout/returns": -7.450580596923828e-09,
      "rollout/raw_reward": 0.4359903381642512,
      "rollout/kl": 0,
      "rollout/rewards": -0.14941632243746145,
      "rollout/rollout_log_probs": -0.17794599384069443,
      "rollout/total_lengths": 8165.251207729469,
      "rollout/advantages": -7.450580596923828e-09,
      "rollout/response_lengths": 311.92270531400965,
      "rollout/truncated": 0.19444444444444445,
      "perf/actor_train_tok_per_s": 13419.081144494183,
      "perf/sleep_time": 12.806060075759888,
      "perf/data_preprocess_time": 0.11535072326660156,
      "perf/wait_time_ratio": 0.34311143625679574,
      "perf/log_probs_time": 133.74723076820374,
      "perf/train_time": 638.6231067180634,
      "perf/wake_up_time": 7.215295076370239,
      "perf/train_wait_time": 333.5708725452423,
      "perf/actor_train_time": 503.82197761535645,
      "perf/step_time": 972.1939792633057,
      "perf/actor_train_tflops": 103.1221275157824,
      "perf/log_probs_tflops": 129.48603103154483,
      "perf/update_weights_time": 1.716897964477539
    },
    {
      "step": 6,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 251.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 5,
      "rollout/shopsim/strict_reward_mean": 0.7318948412698413,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2777777777777778,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.2777777777777778,
      "rollout/shopsim/full_skill_retry_rate": 0,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.7222222222222222,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.23809523809523808,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.3333333333333333,
      "rollout/truncated_ratio": 0.1032258064516129,
      "rollout/step": 6,
      "rollout/shopsim/reward_mean": 0.7318948412698413,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 10.166804044032059,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 68,
      "rollout/response_len/mean": 314.0838709677419,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.1111111111111111,
      "perf/rollout_time": 161.6043736934662,
      "rollout/shopsim/success_rate": 0.7261904761904762,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 10.166804044032059,
      "rollout/shopsim/zero_std_group_drop_count": 5,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.5,
      "rollout/shopsim/group_count": 21,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6111111111111112,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.16666666666666666,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.8888888888888888,
      "perf/tokens_per_gpu_per_sec": 301.2480348603851,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 168,
      "rollout/response_len/max": 1643,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.14285714285714285,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.4444444444444444,
      "perf/effective_tokens_per_gpu_per_sec": 301.2480348603851,
      "rollout/log_probs": -0.1854403093457222,
      "rollout/returns": 7.450580596923828e-09,
      "rollout/raw_reward": 0.5768481182795698,
      "rollout/kl": 0,
      "rollout/rewards": -0.15200128882162034,
      "rollout/rollout_log_probs": -0.18465670198202133,
      "rollout/total_lengths": 6358.995161290322,
      "rollout/advantages": 7.450580596923828e-09,
      "rollout/response_lengths": 314.0838709677419,
      "rollout/truncated": 0.1032258064516129,
      "perf/actor_train_tok_per_s": 13593.931344994806,
      "perf/sleep_time": 12.84454870223999,
      "perf/data_preprocess_time": 0.08673644065856934,
      "perf/wait_time_ratio": 0.33438673351714066,
      "perf/log_probs_time": 76.97732043266296,
      "perf/train_time": 367.8185646533966,
      "perf/wake_up_time": 7.271016836166382,
      "perf/train_wait_time": 184.78244733810425,
      "perf/actor_train_time": 290.0247838497162,
      "perf/step_time": 552.6010119915009,
      "perf/actor_train_tflops": 98.42216838837605,
      "perf/log_probs_tflops": 123.60726489489925,
      "perf/update_weights_time": 1.7390408515930176
    },
    {
      "step": 7,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 233,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 14,
      "rollout/shopsim/strict_reward_mean": 0.7708333333333334,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.30434782608695654,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.391304347826087,
      "rollout/shopsim/full_skill_retry_rate": 0.06666666666666667,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.4782608695652174,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.4666666666666667,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.2608695652173913,
      "rollout/truncated_ratio": 0.05790960451977401,
      "rollout/step": 7,
      "rollout/shopsim/reward_mean": 0.7708333333333334,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 10.202028070580793,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 47,
      "rollout/response_len/mean": 284.47457627118644,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.391304347826087,
      "perf/rollout_time": 214.17310214042664,
      "rollout/shopsim/success_rate": 0.7666666666666667,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.06666666666666667,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.06666666666666667,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 10.202028070580793,
      "rollout/shopsim/zero_std_group_drop_count": 14,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.782608695652174,
      "rollout/shopsim/group_count": 30,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7391304347826086,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.30434782608695654,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.4782608695652174,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 1,
      "perf/tokens_per_gpu_per_sec": 235.09955030200646,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 240,
      "rollout/response_len/max": 2185,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.23333333333333334,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.43478260869565216,
      "perf/effective_tokens_per_gpu_per_sec": 235.09955030200646,
      "rollout/log_probs": -0.19231530278921127,
      "rollout/returns": 0,
      "rollout/raw_reward": 0.6384180790960452,
      "rollout/kl": 0,
      "rollout/rewards": -0.10647217908316413,
      "rollout/rollout_log_probs": -0.1916174367070198,
      "rollout/total_lengths": 6833.632768361582,
      "rollout/advantages": 0,
      "rollout/response_lengths": 284.47457627118644,
      "rollout/truncated": 0.05790960451977401,
      "perf/actor_train_tok_per_s": 13515.574324956322,
      "perf/sleep_time": 12.566070318222046,
      "perf/data_preprocess_time": 0.0993814468383789,
      "perf/wait_time_ratio": 0.34377412285294234,
      "perf/log_probs_time": 94.32879614830017,
      "perf/train_time": 452.84996819496155,
      "perf/wake_up_time": 7.283186912536621,
      "perf/train_wait_time": 237.232492685318,
      "perf/actor_train_time": 357.97309708595276,
      "perf/step_time": 690.0824608802795,
      "perf/actor_train_tflops": 99.06928203313376,
      "perf/log_probs_tflops": 125.32100186998127,
      "perf/update_weights_time": 1.7101953029632568
    },
    {
      "step": 8,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 197,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.57421875,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.1111111111111111,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.37037037037037035,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5185185185185185,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.37037037037037035,
      "rollout/truncated_ratio": 0.30221703617269546,
      "rollout/step": 8,
      "rollout/shopsim/reward_mean": 0.57421875,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 4.482839573560717,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 45,
      "rollout/response_len/mean": 260.51458576429405,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.2962962962962963,
      "perf/rollout_time": 295.12543964385986,
      "rollout/shopsim/success_rate": 0.57421875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 4.482839573560717,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.444444444444445,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5555555555555556,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.2222222222222222,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.4074074074074074,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9259259259259259,
      "perf/tokens_per_gpu_per_sec": 189.12381822236193,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 1323,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.6666666666666666,
      "perf/effective_tokens_per_gpu_per_sec": 189.12381822236193,
      "rollout/log_probs": -0.19516681134700775,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/raw_reward": 0.30921820303383896,
      "rollout/kl": 0,
      "rollout/rewards": -0.014657780058603721,
      "rollout/rollout_log_probs": -0.1943315789103508,
      "rollout/total_lengths": 8711.091015169195,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/response_lengths": 260.51458576429405,
      "rollout/truncated": 0.30221703617269546,
      "perf/actor_train_tok_per_s": 13331.484522764127,
      "perf/sleep_time": 12.51859712600708,
      "perf/data_preprocess_time": 0.11943745613098145,
      "perf/wait_time_ratio": 0.3098662777721318,
      "perf/log_probs_time": 147.38830542564392,
      "perf/train_time": 708.109414100647,
      "perf/wake_up_time": 6.971951007843018,
      "perf/train_wait_time": 317.93726539611816,
      "perf/actor_train_time": 559.9830226898193,
      "perf/step_time": 1026.0466794967651,
      "perf/actor_train_tflops": 105.57641427264927,
      "perf/log_probs_tflops": 133.70802931846072,
      "perf/update_weights_time": 1.703934907913208
    },
    {
      "step": 9,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 203.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 15,
      "rollout/shopsim/strict_reward_mean": 0.6127016129032258,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.16129032258064516,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5384615384615384,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.4838709677419355,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.23076923076923078,
      "rollout/truncated_ratio": 0.2647058823529412,
      "rollout/step": 9,
      "rollout/shopsim/reward_mean": 0.6127016129032258,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 6.311533260847008,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 10,
      "rollout/response_len/mean": 278.8602941176471,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.34615384615384615,
      "perf/rollout_time": 224.8265106678009,
      "rollout/shopsim/success_rate": 0.6088709677419355,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.16129032258064516,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.16129032258064516,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 6.311533260847008,
      "rollout/shopsim/zero_std_group_drop_count": 15,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.769230769230769,
      "rollout/shopsim/group_count": 31,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5769230769230769,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.19230769230769232,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6923076923076923,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9615384615384616,
      "perf/tokens_per_gpu_per_sec": 253.0284343738084,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 248,
      "rollout/response_len/max": 1419,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.16129032258064516,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.34615384615384615,
      "perf/effective_tokens_per_gpu_per_sec": 253.0284343738084,
      "rollout/log_probs": -0.17516595125198364,
      "rollout/returns": -7.450580596923828e-09,
      "rollout/raw_reward": 0.37066993464052284,
      "rollout/kl": 0,
      "rollout/rewards": -0.15117569603756362,
      "rollout/rollout_log_probs": -0.17387615889310837,
      "rollout/total_lengths": 8103.524509803921,
      "rollout/advantages": -7.450580596923828e-09,
      "rollout/response_lengths": 278.8602941176471,
      "rollout/truncated": 0.2647058823529412,
      "perf/actor_train_tok_per_s": 13363.047306224207,
      "perf/sleep_time": 12.554746389389038,
      "perf/data_preprocess_time": 0.1154632568359375,
      "perf/wait_time_ratio": 0.28355729301056776,
      "perf/log_probs_time": 130.5078706741333,
      "perf/train_time": 626.1998240947723,
      "perf/wake_up_time": 7.15744948387146,
      "perf/train_wait_time": 247.84051156044006,
      "perf/actor_train_time": 494.83294105529785,
      "perf/step_time": 874.0403356552124,
      "perf/actor_train_tflops": 103.55064242672432,
      "perf/log_probs_tflops": 130.87401976984708,
      "perf/update_weights_time": 1.7218284606933594
    },
    {
      "step": 10,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 218.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.6733668154761905,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.34615384615384615,
      "rollout/shopsim/full_skill_retry_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.6153846153846154,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.2692307692307692,
      "rollout/truncated_ratio": 0.05014749262536873,
      "rollout/step": 10,
      "rollout/shopsim/reward_mean": 0.6733668154761905,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "rollout/zero_std/count_0.7": 1,
      "perf/longest_effective_sample_tokens_per_sec": 22.10669119898411,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 59,
      "rollout/response_len/mean": 292.3200589970501,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.23076923076923078,
      "perf/rollout_time": 185.28326845169067,
      "rollout/shopsim/success_rate": 0.6328125,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.1875,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 22.10669119898411,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.884615384615385,
      "rollout/shopsim/group_count": 32,
      "rollout/zero_std/count_1.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5384615384615384,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.5,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6153846153846154,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9615384615384616,
      "perf/tokens_per_gpu_per_sec": 267.4189116699376,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5769230769230769,
      "perf/effective_tokens_per_gpu_per_sec": 267.4189116699376,
      "rollout/log_probs": -0.17175283282995224,
      "rollout/returns": 0.003725258633494377,
      "rollout/raw_reward": 0.5326309874982441,
      "rollout/kl": 0,
      "rollout/rewards": -0.012942746237667613,
      "rollout/rollout_log_probs": -0.17094164341688156,
      "rollout/total_lengths": 6124.069321533923,
      "rollout/advantages": 0.003725258633494377,
      "rollout/response_lengths": 292.3200589970501,
      "rollout/truncated": 0.05014749262536873,
      "perf/actor_train_tok_per_s": 13630.84679488102,
      "perf/sleep_time": 12.728774785995483,
      "perf/data_preprocess_time": 0.09020853042602539,
      "perf/wait_time_ratio": 0.3512057229901636,
      "perf/log_probs_time": 79.30752444267273,
      "perf/train_time": 384.70793986320496,
      "perf/wake_up_time": 7.076560974121094,
      "perf/train_wait_time": 208.250342130661,
      "perf/actor_train_time": 304.61196303367615,
      "perf/step_time": 592.958281993866,
      "perf/actor_train_tflops": 95.57724300996948,
      "perf/log_probs_tflops": 122.36742065863066,
      "perf/update_weights_time": 1.730684757232666
    },
    {
      "step": 11,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 193,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 13,
      "rollout/shopsim/strict_reward_mean": 0.6905659893267652,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.14285714285714285,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.47619047619047616,
      "rollout/shopsim/full_skill_retry_rate": 0.13793103448275862,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5714285714285714,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.4482758620689655,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.23809523809523808,
      "rollout/truncated_ratio": 0.12768817204301075,
      "rollout/step": 11,
      "rollout/shopsim/reward_mean": 0.6905659893267652,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 5.222839727531851,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 55,
      "rollout/response_len/mean": 240.18682795698925,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.14285714285714285,
      "perf/rollout_time": 234.73820066452026,
      "rollout/shopsim/success_rate": 0.6594827586206896,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.10344827586206896,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.13793103448275862,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 5.222839727531851,
      "rollout/shopsim/zero_std_group_drop_count": 13,
      "rollout/shopsim/full_skill_retry_success_rate": 0.25,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.190476190476191,
      "rollout/shopsim/group_count": 29,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.47619047619047616,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.38095238095238093,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.42857142857142855,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9523809523809523,
      "perf/tokens_per_gpu_per_sec": 190.31734022638952,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 232,
      "rollout/response_len/max": 1226,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0.034482758620689655,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.27586206896551724,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.38095238095238093,
      "perf/effective_tokens_per_gpu_per_sec": 190.31734022638952,
      "rollout/log_probs": -0.18521509319543839,
      "rollout/returns": 1.5133991837501526e-09,
      "rollout/raw_reward": 0.6136792754736303,
      "rollout/kl": 0,
      "rollout/rewards": -0.11667181627564532,
      "rollout/rollout_log_probs": -0.18444393575191498,
      "rollout/total_lengths": 6906.643817204301,
      "rollout/advantages": 1.5133991837501526e-09,
      "rollout/response_lengths": 240.18682795698925,
      "rollout/truncated": 0.12768817204301075,
      "perf/actor_train_tok_per_s": 13465.266235992629,
      "perf/sleep_time": 12.684320449829102,
      "perf/data_preprocess_time": 0.09803628921508789,
      "perf/wait_time_ratio": 0.3486815779341211,
      "perf/log_probs_time": 99.0292444229126,
      "perf/train_time": 481.42845487594604,
      "perf/wake_up_time": 6.948862552642822,
      "perf/train_wait_time": 257.73143768310547,
      "perf/actor_train_time": 381.6146602630615,
      "perf/step_time": 739.1598925590515,
      "perf/actor_train_tflops": 99.86453016655797,
      "perf/log_probs_tflops": 128.2778266626316,
      "perf/update_weights_time": 1.755722999572754
    },
    {
      "step": 12,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 216,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.68505859375,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.125,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.75,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.4583333333333333,
      "rollout/truncated_ratio": 0.04353741496598639,
      "rollout/step": 12,
      "rollout/shopsim/reward_mean": 0.68505859375,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 19.388071952237873,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 68,
      "rollout/response_len/mean": 291.7646258503401,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.25,
      "perf/rollout_time": 211.26391577720642,
      "rollout/shopsim/success_rate": 0.66796875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 19.388071952237873,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0.2,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.833333333333333,
      "rollout/shopsim/group_count": 32,
      "rollout/zero_std/count_1.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5416666666666666,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.4166666666666667,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5416666666666666,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.875,
      "perf/tokens_per_gpu_per_sec": 253.76671545053438,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/internalization_deficit_rate": 0.03125,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.375,
      "perf/effective_tokens_per_gpu_per_sec": 253.76671545053438,
      "rollout/log_probs": -0.17173174023628235,
      "rollout/returns": -6.51925802230835e-09,
      "rollout/raw_reward": 0.4006462585034014,
      "rollout/kl": 0,
      "rollout/rewards": -0.039904331268907405,
      "rollout/rollout_log_probs": -0.17078875750303268,
      "rollout/total_lengths": 7119.885714285714,
      "rollout/advantages": -6.51925802230835e-09,
      "rollout/response_lengths": 291.7646258503401,
      "rollout/truncated": 0.04353741496598639,
      "perf/actor_train_tok_per_s": 13516.72868714718,
      "perf/sleep_time": 12.931987285614014,
      "perf/data_preprocess_time": 0.09810876846313477,
      "perf/wait_time_ratio": 0.3240829745821438,
      "perf/log_probs_time": 101.39136910438538,
      "perf/train_time": 489.3509018421173,
      "perf/wake_up_time": 7.187591314315796,
      "perf/train_wait_time": 234.62982869148254,
      "perf/actor_train_time": 387.1584701538086,
      "perf/step_time": 723.9807305335999,
      "perf/actor_train_tflops": 99.1875739146668,
      "perf/log_probs_tflops": 126.24779181330823,
      "perf/update_weights_time": 1.7580935955047607
    },
    {
      "step": 13,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 202,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.75,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.10714285714285714,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5357142857142857,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5357142857142857,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.25,
      "rollout/truncated_ratio": 0,
      "rollout/step": 13,
      "rollout/shopsim/reward_mean": 0.75,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 6.982283570536215,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 74,
      "rollout/response_len/mean": 244.89528795811518,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.10714285714285714,
      "perf/rollout_time": 161.26529216766357,
      "rollout/shopsim/success_rate": 0.734375,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/zero_std/count_0.5": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 6.982283570536215,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.392857142857143,
      "rollout/shopsim/group_count": 32,
      "rollout/zero_std/count_1.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7142857142857143,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.5,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.35714285714285715,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.8571428571428571,
      "perf/tokens_per_gpu_per_sec": 217.53750933292505,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 1126,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.42857142857142855,
      "perf/effective_tokens_per_gpu_per_sec": 217.53750933292505,
      "rollout/log_probs": -0.1779138743877411,
      "rollout/returns": -3.725290298461914e-09,
      "rollout/raw_reward": 0.5706806282722513,
      "rollout/kl": 0,
      "rollout/rewards": 0.04672105373720432,
      "rollout/rollout_log_probs": -0.17717094719409943,
      "rollout/total_lengths": 5217.806282722513,
      "rollout/advantages": -3.725290298461914e-09,
      "rollout/response_lengths": 244.89528795811518,
      "rollout/truncated": 0,
      "perf/actor_train_tok_per_s": 13941.810343099945,
      "perf/sleep_time": 12.94321346282959,
      "perf/data_preprocess_time": 0.0705862045288086,
      "perf/wait_time_ratio": 0.40470731282457595,
      "perf/log_probs_time": 55.687525272369385,
      "perf/train_time": 271.0225977897644,
      "perf/wake_up_time": 7.117730140686035,
      "perf/train_wait_time": 184.25361108779907,
      "perf/actor_train_time": 214.44869256019592,
      "perf/step_time": 455.2762088775635,
      "perf/actor_train_tflops": 92.76463891772502,
      "perf/log_probs_tflops": 119.07667192623052,
      "perf/update_weights_time": 1.7111845016479492
    },
    {
      "step": 14,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 205.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.6458333333333334,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.36,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.52,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.16,
      "rollout/truncated_ratio": 0.09023668639053255,
      "rollout/step": 14,
      "rollout/shopsim/reward_mean": 0.6458333333333334,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 18.181009220225096,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 58,
      "rollout/response_len/mean": 276.2174556213018,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.04,
      "perf/rollout_time": 225.29002380371094,
      "rollout/shopsim/success_rate": 0.6171875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 18.181009220225096,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.2,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.68,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.48,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.36,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.92,
      "perf/tokens_per_gpu_per_sec": 207.20291654224187,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.48,
      "perf/effective_tokens_per_gpu_per_sec": 207.20291654224187,
      "rollout/log_probs": -0.16990060359239578,
      "rollout/returns": -2.2351741790771484e-08,
      "rollout/raw_reward": 0.2948717948717949,
      "rollout/kl": 0,
      "rollout/rewards": -0.07430439768634604,
      "rollout/rollout_log_probs": -0.16935279965400696,
      "rollout/total_lengths": 6776.600591715976,
      "rollout/advantages": -2.2351741790771484e-08,
      "rollout/response_lengths": 276.2174556213018,
      "rollout/truncated": 0.09023668639053255,
      "perf/actor_train_tok_per_s": 13654.305638556774,
      "perf/sleep_time": 12.856248140335083,
      "perf/data_preprocess_time": 0.09143424034118652,
      "perf/wait_time_ratio": 0.3677627243166203,
      "perf/log_probs_time": 90.69216537475586,
      "perf/train_time": 427.09525179862976,
      "perf/wake_up_time": 7.194684743881226,
      "perf/train_wait_time": 248.43475604057312,
      "perf/actor_train_time": 335.4972505569458,
      "perf/step_time": 675.5300078392029,
      "perf/actor_train_tflops": 100.99217529904499,
      "perf/log_probs_tflops": 124.53334861060263,
      "perf/update_weights_time": 1.666468858718872
    },
    {
      "step": 15,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 198,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 13,
      "rollout/shopsim/strict_reward_mean": 0.684985632183908,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.125,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.06896551724137931,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5833333333333334,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.4482758620689655,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.20833333333333334,
      "rollout/truncated_ratio": 0.06073446327683616,
      "rollout/step": 15,
      "rollout/shopsim/reward_mean": 0.684985632183908,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 9.73296432805524,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 52,
      "rollout/response_len/mean": 252.37853107344634,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.16666666666666666,
      "perf/rollout_time": 203.74060082435608,
      "rollout/shopsim/success_rate": 0.6724137931034483,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.06896551724137931,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.06896551724137931,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 9.73296432805524,
      "rollout/shopsim/zero_std_group_drop_count": 13,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.666666666666667,
      "rollout/shopsim/group_count": 29,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.375,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.4583333333333333,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9166666666666666,
      "perf/tokens_per_gpu_per_sec": 219.25428618182335,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 232,
      "rollout/response_len/max": 1983,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.1724137931034483,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.6666666666666666,
      "perf/effective_tokens_per_gpu_per_sec": 219.25428618182335,
      "rollout/log_probs": -0.17899398505687714,
      "rollout/returns": -5.587935447692871e-09,
      "rollout/raw_reward": 0.5144774011299436,
      "rollout/kl": 0,
      "rollout/rewards": -0.0633846496121358,
      "rollout/rollout_log_probs": -0.1782628893852234,
      "rollout/total_lengths": 6714.526836158192,
      "rollout/advantages": -5.587935447692871e-09,
      "rollout/response_lengths": 252.37853107344634,
      "rollout/truncated": 0.06073446327683616,
      "perf/actor_train_tok_per_s": 13482.656413145567,
      "perf/sleep_time": 12.26793122291565,
      "perf/data_preprocess_time": 0.09775280952453613,
      "perf/wait_time_ratio": 0.3368840023088133,
      "perf/log_probs_time": 91.2160484790802,
      "perf/train_time": 445.44486021995544,
      "perf/wake_up_time": 7.183570146560669,
      "perf/train_wait_time": 226.30014634132385,
      "perf/actor_train_time": 352.5926089286804,
      "perf/step_time": 671.7450065612793,
      "perf/actor_train_tflops": 98.92585431815282,
      "perf/log_probs_tflops": 127.46486919837608,
      "perf/update_weights_time": 1.7215125560760498
    },
    {
      "step": 16,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 221,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.7890625,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.52,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.32,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.24,
      "rollout/truncated_ratio": 0.01957585644371941,
      "rollout/step": 16,
      "rollout/shopsim/reward_mean": 0.7890625,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 11.25939735540376,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 57,
      "rollout/response_len/mean": 256.0391517128874,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.32,
      "perf/rollout_time": 171.14592719078064,
      "rollout/shopsim/success_rate": 0.7890625,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 11.25939735540376,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.8,
      "rollout/shopsim/group_count": 32,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.68,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.44,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.96,
      "perf/tokens_per_gpu_per_sec": 229.2663380546615,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 1927,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.52,
      "perf/effective_tokens_per_gpu_per_sec": 229.2663380546615,
      "rollout/log_probs": -0.1791928932070732,
      "rollout/returns": 7.450580596923828e-09,
      "rollout/raw_reward": 0.5138662316476346,
      "rollout/kl": 0,
      "rollout/rewards": -0.05429584706782517,
      "rollout/rollout_log_probs": -0.17832156270742416,
      "rollout/total_lengths": 5847.324632952692,
      "rollout/advantages": 7.450580596923828e-09,
      "rollout/response_lengths": 256.0391517128874,
      "rollout/truncated": 0.01957585644371941,
      "perf/actor_train_tok_per_s": 13697.26272728488,
      "perf/sleep_time": 12.923967838287354,
      "perf/data_preprocess_time": 0.08007431030273438,
      "perf/wait_time_ratio": 0.3694387410397893,
      "perf/log_probs_time": 69.07073712348938,
      "perf/train_time": 331.5771436691284,
      "perf/wake_up_time": 7.070033550262451,
      "perf/train_wait_time": 194.2673146724701,
      "perf/actor_train_time": 261.6880519390106,
      "perf/step_time": 525.8444583415985,
      "perf/actor_train_tflops": 95.75586705850814,
      "perf/log_probs_tflops": 120.9299690715195,
      "perf/update_weights_time": 1.727674961090088
    },
    {
      "step": 17,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 250.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 15,
      "rollout/shopsim/strict_reward_mean": 0.7872983870967742,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.16,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.32,
      "rollout/shopsim/full_skill_retry_rate": 0.06451612903225806,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.64,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.4838709677419355,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.32,
      "rollout/truncated_ratio": 0,
      "rollout/step": 17,
      "rollout/shopsim/reward_mean": 0.7872983870967742,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 6.58591922252755,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 61,
      "rollout/response_len/mean": 301.03177257525084,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.2,
      "perf/rollout_time": 170.05978393554688,
      "rollout/shopsim/success_rate": 0.782258064516129,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.06451612903225806,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.06451612903225806,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 6.58591922252755,
      "rollout/shopsim/zero_std_group_drop_count": 15,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.56,
      "rollout/shopsim/group_count": 31,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.72,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.36,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.92,
      "perf/tokens_per_gpu_per_sec": 264.63781711646027,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 248,
      "rollout/response_len/max": 1120,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.1935483870967742,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.52,
      "perf/effective_tokens_per_gpu_per_sec": 264.63781711646027,
      "rollout/log_probs": -0.16822337359189987,
      "rollout/returns": -1.1175870895385742e-08,
      "rollout/raw_reward": 0.6396321070234113,
      "rollout/kl": 0,
      "rollout/rewards": -0.03372907857831107,
      "rollout/rollout_log_probs": -0.16753577440977097,
      "rollout/total_lengths": 5838.444816053512,
      "rollout/advantages": -1.1175870895385742e-08,
      "rollout/response_lengths": 301.03177257525084,
      "rollout/truncated": 0,
      "perf/actor_train_tok_per_s": 13662.433983906361,
      "perf/sleep_time": 12.824751853942871,
      "perf/data_preprocess_time": 0.08175420761108398,
      "perf/wait_time_ratio": 0.374205340678814,
      "perf/log_probs_time": 66.40709257125854,
      "perf/train_time": 322.76292538642883,
      "perf/wake_up_time": 7.065443992614746,
      "perf/train_wait_time": 193.00198340415955,
      "perf/actor_train_time": 255.54670596122742,
      "perf/step_time": 515.7649087905884,
      "perf/actor_train_tflops": 94.08390680193634,
      "perf/log_probs_tflops": 120.68405916832172,
      "perf/update_weights_time": 1.741757869720459
    },
    {
      "step": 18,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 225,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.5879603794642857,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.35714285714285715,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5357142857142857,
      "rollout/shopsim/full_skill_retry_rate": 0.34375,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.42857142857142855,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.2857142857142857,
      "rollout/truncated_ratio": 0,
      "rollout/step": 18,
      "rollout/shopsim/reward_mean": 0.5879603794642857,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 4.560981734509255,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 57,
      "rollout/response_len/mean": 269.37282780410743,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.17857142857142858,
      "perf/rollout_time": 242.0531508922577,
      "rollout/shopsim/success_rate": 0.57421875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.34375,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.34375,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 4.560981734509255,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.714285714285714,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.75,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.25,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.42857142857142855,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.8571428571428571,
      "perf/tokens_per_gpu_per_sec": 176.11111378994036,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 1104,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 6,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.6428571428571429,
      "perf/effective_tokens_per_gpu_per_sec": 176.11111378994036,
      "rollout/log_probs": -0.18028194457292557,
      "rollout/returns": -3.725290298461914e-09,
      "rollout/raw_reward": 0.34509704355675924,
      "rollout/kl": 0,
      "rollout/rewards": 0.012306475761754065,
      "rollout/rollout_log_probs": -0.179410919547081,
      "rollout/total_lengths": 5720.415481832543,
      "rollout/advantages": -3.725290298461914e-09,
      "rollout/response_lengths": 269.37282780410743,
      "rollout/truncated": 0,
      "perf/actor_train_tok_per_s": 13686.925149084824,
      "perf/sleep_time": 12.656147003173828,
      "perf/data_preprocess_time": 0.0811319351196289,
      "perf/wait_time_ratio": 0.4411596590061575,
      "perf/log_probs_time": 70.07508420944214,
      "perf/train_time": 335.51728653907776,
      "perf/wake_up_time": 7.035004138946533,
      "perf/train_wait_time": 264.8640065193176,
      "perf/actor_train_time": 264.56073665618896,
      "perf/step_time": 600.3812930583954,
      "perf/actor_train_tflops": 94.66658655504048,
      "perf/log_probs_tflops": 119.13441207754188,
      "perf/update_weights_time": 1.727344036102295
    },
    {
      "step": 19,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 220,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.8119436553030303,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.16666666666666666,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5416666666666666,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.5416666666666666,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.25,
      "rollout/truncated_ratio": 0.029850746268656716,
      "rollout/step": 19,
      "rollout/shopsim/reward_mean": 0.8119436553030303,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 16.125557997246652,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 75,
      "rollout/response_len/mean": 280.3626865671642,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.08333333333333333,
      "perf/rollout_time": 254.00671410560608,
      "rollout/shopsim/success_rate": 0.8046875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 16.125557997246652,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.166666666666667,
      "rollout/shopsim/group_count": 32,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5416666666666666,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.20833333333333334,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.9166666666666666,
      "perf/tokens_per_gpu_per_sec": 184.87995549785174,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.4166666666666667,
      "perf/effective_tokens_per_gpu_per_sec": 184.87995549785174,
      "rollout/log_probs": -0.1604425087571144,
      "rollout/returns": 9.313225746154785e-09,
      "rollout/raw_reward": 0.5381049298959746,
      "rollout/kl": 0,
      "rollout/rewards": -0.07268341085804042,
      "rollout/rollout_log_probs": -0.15949886292219162,
      "rollout/total_lengths": 5700.028358208955,
      "rollout/advantages": 9.313225746154785e-09,
      "rollout/response_lengths": 280.3626865671642,
      "rollout/truncated": 0.029850746268656716,
      "perf/actor_train_tok_per_s": 13791.47203448624,
      "perf/sleep_time": 12.428958892822266,
      "perf/data_preprocess_time": 0.08535099029541016,
      "perf/wait_time_ratio": 0.44060997653307293,
      "perf/log_probs_time": 73.25257682800293,
      "perf/train_time": 351.1751217842102,
      "perf/wake_up_time": 7.1393821239471436,
      "perf/train_wait_time": 276.60711789131165,
      "perf/actor_train_time": 276.9116299152374,
      "perf/step_time": 627.7822396755219,
      "perf/actor_train_tflops": 94.77852463718025,
      "perf/log_probs_tflops": 119.42822525714459,
      "perf/update_weights_time": 1.7195663452148438
    },
    {
      "step": 20,
      "round": 0,
      "run_id": "8ov14yyc",
      "rollout/response_len/median": 228.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/strict_reward_mean": 0.72607421875,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/target_rho": 0.38162078131953847,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/rho_mean": 0.44903167489954565,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/target_rho": 0.26310000301127795,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/inclusion_rate": 0.46153846153846156,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-4d958fc1d69e/inclusion_rate": 0.23076923076923078,
      "rollout/truncated_ratio": 0.10698689956331878,
      "rollout/step": 20,
      "rollout/shopsim/reward_mean": 0.72607421875,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/target_rho": 0.8999999999999999,
      "perf/longest_effective_sample_tokens_per_sec": 4.7374021152009735,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/target_rho": 0.2002050350677727,
      "rollout/response_len/min": 66,
      "rollout/response_len/mean": 288.96615720524017,
      "rollout/shopsim/chunk/chunk-a71a5634ee99/inclusion_rate": 0.11538461538461539,
      "perf/rollout_time": 296.99822092056274,
      "rollout/shopsim/success_rate": 0.71875,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "rollout/shopsim/chunk/chunk-d3a9ac5af7ef/target_rho": 0.5174498242544147,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.49059913361494567,
      "perf/longest_sample_tokens_per_sec": 4.7374021152009735,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4733805357216591,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.346153846153846,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6153846153846154,
      "rollout/shopsim/chunk/chunk-b64b0567ce98/inclusion_rate": 0.4230769230769231,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6153846153846154,
      "rollout/shopsim/chunk/chunk-f58e352cdad0/inclusion_rate": 0.8461538461538461,
      "perf/tokens_per_gpu_per_sec": 222.8068902059153,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.2001840055744324,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/response_len/max": 1407,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.5796673375052258,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.48411009292619,
      "rollout/shopsim/rho_min": 0.2001840055744324,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.3076923076923077,
      "perf/effective_tokens_per_gpu_per_sec": 222.8068902059153,
      "rollout/log_probs": -0.15676091611385345,
      "rollout/returns": -4.190951585769653e-09,
      "rollout/raw_reward": 0.39042030567685587,
      "rollout/kl": 0,
      "rollout/rewards": -0.32754344434587196,
      "rollout/rollout_log_probs": -0.1560058742761612,
      "rollout/total_lengths": 7647.174672489083,
      "rollout/advantages": -4.190951585769653e-09,
      "rollout/response_lengths": 288.96615720524017,
      "rollout/truncated": 0.10698689956331878,
      "perf/actor_train_tok_per_s": 13503.93021286861,
      "perf/sleep_time": 12.33907437324524,
      "perf/data_preprocess_time": 0.13480496406555176,
      "perf/wait_time_ratio": 0.32782441889151076,
      "perf/log_probs_time": 138.17161464691162,
      "perf/train_time": 657.5582048892975,
      "perf/wake_up_time": 7.253126859664917,
      "perf/train_wait_time": 320.6954290866852,
      "perf/actor_train_time": 518.7239484786987,
      "perf/step_time": 978.2536339759827,
      "perf/actor_train_tflops": 100.93821669967541,
      "perf/log_probs_tflops": 126.31410209845265,
      "perf/update_weights_time": 1.7155601978302002
    },
    {
      "step": 21,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 248.88871944362396,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8666666666666667,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5333333333333333,
      "perf/longest_effective_sample_tokens_per_sec": 15.587177879318743,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/reward_mean": 0.8606770833333334,
      "rollout/shopsim/success_rate": 0.85546875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6333333333333333,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.4666666666666667,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 27,
      "rollout/response_len/median": 212,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.0625,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.43333333333333335,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 21,
      "perf/longest_sample_tokens_per_sec": 15.587177879318743,
      "rollout/repetition_frac": 0.001002004008016032,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "perf/effective_tokens_per_gpu_per_sec": 248.88871944362396,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5666666666666667,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.8606770833333334,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.166666666666667,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.17835671342685372,
      "perf/rollout_time": 262.7800896167755,
      "rollout/response_len/mean": 262.1362725450902,
      "rollout/total_lengths": 8698.97394789579,
      "rollout/raw_reward": 0.458249832999332,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.16026732325553894,
      "rollout/truncated": 0.17835671342685372,
      "rollout/log_probs": -0.1615421399474144,
      "rollout/rewards": -0.34615140729294513,
      "rollout/returns": 2.3283064365386963e-09,
      "rollout/advantages": 2.3283064365386963e-09,
      "rollout/response_lengths": 262.1362725450902,
      "perf/log_probs_time": 176.04646611213684,
      "perf/actor_train_tok_per_s": 13379.245151437863,
      "perf/train_time": 825.6104173660278,
      "perf/actor_train_tflops": 104.27603772360105,
      "perf/wake_up_time": 4.064673185348511,
      "perf/train_wait_time": 269.97549962997437,
      "perf/actor_train_time": 648.8838422298431,
      "perf/update_weights_time": 1.7147791385650635,
      "perf/step_time": 1095.5859169960022,
      "perf/sleep_time": 6.644998550415039,
      "perf/data_preprocess_time": 0.1478438377380371,
      "perf/wait_time_ratio": 0.24642111170087222,
      "perf/log_probs_tflops": 128.1158652880783
    },
    {
      "step": 22,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 269.5151262127911,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.52,
      "perf/longest_effective_sample_tokens_per_sec": 7.384656167991301,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 2008,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.56,
      "rollout/shopsim/reward_mean": 0.7371651785714286,
      "rollout/shopsim/success_rate": 0.72265625,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.64,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.48,
      "rollout/shopsim/full_skill_retry_success_rate": 0.5,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 34,
      "rollout/response_len/median": 255,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 22,
      "perf/longest_sample_tokens_per_sec": 7.384656167991301,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "perf/effective_tokens_per_gpu_per_sec": 269.5151262127911,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.52,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/strict_reward_mean": 0.7371651785714286,
      "rollout/shopsim/internalization_deficit_rate": 0.03125,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.8,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.22406639004149378,
      "perf/rollout_time": 271.91516494750977,
      "rollout/response_len/mean": 304.0881742738589,
      "rollout/total_lengths": 8448.86203319502,
      "rollout/raw_reward": 0.4142708950800237,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.16393698751926422,
      "rollout/truncated": 0.22406639004149378,
      "rollout/log_probs": -0.16462386399507523,
      "rollout/rewards": -0.28802968926449535,
      "rollout/returns": 1.6763806343078613e-08,
      "rollout/advantages": 1.6763806343078613e-08,
      "rollout/response_lengths": 304.0881742738589,
      "perf/log_probs_time": 161.11005187034607,
      "perf/actor_train_tok_per_s": 13358.288531288206,
      "perf/train_time": 772.0073940753937,
      "perf/actor_train_tflops": 103.6037547721573,
      "perf/wake_up_time": 6.5107526779174805,
      "perf/train_wait_time": 294.09449458122253,
      "perf/actor_train_time": 609.711564540863,
      "perf/update_weights_time": 1.7182445526123047,
      "perf/step_time": 1066.1018886566162,
      "perf/sleep_time": 11.182740688323975,
      "perf/data_preprocess_time": 0.1449124813079834,
      "perf/wait_time_ratio": 0.27585965066791873,
      "perf/log_probs_tflops": 130.69411598078509
    },
    {
      "step": 23,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 216.470219797668,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9666666666666667,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.6,
      "perf/longest_effective_sample_tokens_per_sec": 4.077226566646936,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1343,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6,
      "rollout/shopsim/reward_mean": 0.7799479166666666,
      "rollout/shopsim/success_rate": 0.77734375,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.8,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.36666666666666664,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 33,
      "rollout/response_len/median": 210,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.0625,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.43333333333333335,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 23,
      "perf/longest_sample_tokens_per_sec": 4.077226566646936,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "perf/effective_tokens_per_gpu_per_sec": 216.470219797668,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5333333333333333,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.7799479166666666,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.3,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.3736654804270463,
      "perf/rollout_time": 329.39057421684265,
      "rollout/response_len/mean": 253.7482206405694,
      "rollout/total_lengths": 9835.615658362989,
      "rollout/raw_reward": 0.3140569395017794,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.16511093825101852,
      "rollout/truncated": 0.3736654804270463,
      "rollout/log_probs": -0.16599081456661224,
      "rollout/rewards": -0.21075694392902572,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/response_lengths": 253.7482206405694,
      "perf/log_probs_time": 217.60114526748657,
      "perf/actor_train_tok_per_s": 13374.156882269874,
      "perf/train_time": 1045.1599757671356,
      "perf/actor_train_tflops": 108.03482301564249,
      "perf/wake_up_time": 6.550279140472412,
      "perf/train_wait_time": 351.7172546386719,
      "perf/actor_train_time": 826.611508846283,
      "perf/update_weights_time": 1.7253479957580566,
      "perf/step_time": 1396.8772304058075,
      "perf/sleep_time": 10.991825342178345,
      "perf/data_preprocess_time": 0.17418527603149414,
      "perf/wait_time_ratio": 0.25178823663443517,
      "perf/log_probs_tflops": 136.79895534116747
    },
    {
      "step": 24,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 245.28126570189033,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.7857142857142857,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5,
      "perf/longest_effective_sample_tokens_per_sec": 5.677470094589426,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1170,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5357142857142857,
      "rollout/shopsim/reward_mean": 0.8203125,
      "rollout/shopsim/success_rate": 0.8125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6428571428571429,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.32142857142857145,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 78,
      "rollout/response_len/median": 257.5,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.32142857142857145,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 24,
      "perf/longest_sample_tokens_per_sec": 5.677470094589426,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "perf/effective_tokens_per_gpu_per_sec": 245.28126570189033,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/strict_reward_mean": 0.8203125,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.607142857142857,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0,
      "perf/rollout_time": 206.07770371437073,
      "rollout/response_len/mean": 301.7731343283582,
      "rollout/total_lengths": 6070.598507462687,
      "rollout/raw_reward": 0.6059701492537314,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.1606895476579666,
      "rollout/truncated": 0,
      "rollout/log_probs": -0.16139420121908188,
      "rollout/rewards": -0.1253614064028014,
      "rollout/returns": 9.313225746154785e-10,
      "rollout/advantages": 9.313225746154785e-10,
      "rollout/response_lengths": 301.7731343283582,
      "perf/log_probs_time": 78.88224959373474,
      "perf/actor_train_tok_per_s": 13785.36163769729,
      "perf/train_time": 374.66028690338135,
      "perf/actor_train_tflops": 95.9201554104046,
      "perf/wake_up_time": 6.414995431900024,
      "perf/train_wait_time": 228.186283826828,
      "perf/actor_train_time": 295.04492568969727,
      "perf/update_weights_time": 1.7009613513946533,
      "perf/step_time": 602.8465707302094,
      "perf/sleep_time": 11.495723009109497,
      "perf/data_preprocess_time": 0.09526538848876953,
      "perf/wait_time_ratio": 0.3785146916410805,
      "perf/log_probs_tflops": 119.59072022313302
    },
    {
      "step": 25,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 215.8468595911726,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8888888888888888,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.2962962962962963,
      "perf/longest_effective_sample_tokens_per_sec": 3.9428060967390817,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1177,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.7407407407407407,
      "rollout/shopsim/reward_mean": 0.8268229166666666,
      "rollout/shopsim/success_rate": 0.8203125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5555555555555556,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5925925925925926,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 60,
      "rollout/response_len/median": 279.5,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2222222222222222,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 25,
      "perf/longest_sample_tokens_per_sec": 3.9428060967390817,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "perf/effective_tokens_per_gpu_per_sec": 215.8468595911726,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5555555555555556,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/strict_reward_mean": 0.8268229166666666,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.8518518518518516,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.10756501182033097,
      "perf/rollout_time": 298.5183575153351,
      "rollout/response_len/mean": 304.6536643026005,
      "rollout/total_lengths": 7878.430260047281,
      "rollout/raw_reward": 0.5638297872340425,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.16572517901659012,
      "rollout/truncated": 0.10756501182033097,
      "rollout/log_probs": -0.1665768250823021,
      "rollout/rewards": -0.026581946487404213,
      "rollout/returns": 6.05359673500061e-09,
      "rollout/advantages": 6.05359673500061e-09,
      "rollout/response_lengths": 304.6536643026005,
      "perf/log_probs_time": 132.72907328605652,
      "perf/actor_train_tok_per_s": 13441.03321736369,
      "perf/train_time": 629.3947098255157,
      "perf/actor_train_tflops": 102.37545037317308,
      "perf/wake_up_time": 6.4231250286102295,
      "perf/train_wait_time": 320.229195356369,
      "perf/actor_train_time": 495.88092613220215,
      "perf/update_weights_time": 1.7131924629211426,
      "perf/step_time": 949.6239051818848,
      "perf/sleep_time": 10.887051820755005,
      "perf/data_preprocess_time": 0.13532352447509766,
      "perf/wait_time_ratio": 0.3372168640752935,
      "perf/log_probs_tflops": 127.49287423723126
    },
    {
      "step": 26,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 194.93156268062972,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8333333333333334,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.4166666666666667,
      "perf/longest_effective_sample_tokens_per_sec": 3.5583318984510797,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1182,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5833333333333334,
      "rollout/shopsim/reward_mean": 0.8140625,
      "rollout/shopsim/success_rate": 0.8125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7083333333333334,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.3333333333333333,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 59,
      "rollout/response_len/median": 239,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.375,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 26,
      "perf/longest_sample_tokens_per_sec": 3.5583318984510797,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 194.93156268062972,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 4,
      "rollout/shopsim/strict_reward_mean": 0.8140625,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.75,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.32346723044397463,
      "perf/rollout_time": 332.1781198978424,
      "rollout/response_len/mean": 273.79281183932346,
      "rollout/total_lengths": 8941.77378435518,
      "rollout/raw_reward": 0.45898520084566596,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15837904810905457,
      "rollout/truncated": 0.32346723044397463,
      "rollout/log_probs": -0.1589760184288025,
      "rollout/rewards": -0.14051031496131647,
      "rollout/returns": -3.259629011154175e-09,
      "rollout/advantages": -3.259629011154175e-09,
      "rollout/response_lengths": 273.79281183932346,
      "perf/log_probs_time": 169.1874542236328,
      "perf/actor_train_tok_per_s": 13426.282357851402,
      "perf/train_time": 800.033659696579,
      "perf/actor_train_tflops": 106.61712098335836,
      "perf/wake_up_time": 6.54135799407959,
      "perf/train_wait_time": 355.0079185962677,
      "perf/actor_train_time": 630.0268216133118,
      "perf/update_weights_time": 1.720334768295288,
      "perf/step_time": 1155.0415782928467,
      "perf/sleep_time": 11.558203935623169,
      "perf/data_preprocess_time": 0.15378093719482422,
      "perf/wait_time_ratio": 0.3073550989575371,
      "perf/log_probs_tflops": 132.34166045968433
    },
    {
      "step": 27,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 241.75006390833732,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9310344827586207,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.41379310344827586,
      "perf/longest_effective_sample_tokens_per_sec": 9.714366071898715,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 2343,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6551724137931034,
      "rollout/shopsim/reward_mean": 0.921875,
      "rollout/shopsim/success_rate": 0.921875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7586206896551724,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5517241379310345,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 79,
      "rollout/response_len/median": 251.5,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.09375,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4482758620689655,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 27,
      "perf/longest_sample_tokens_per_sec": 9.714366071898715,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "perf/effective_tokens_per_gpu_per_sec": 241.75006390833732,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.6896551724137931,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/zero_std/count_1.0": 9,
      "rollout/shopsim/strict_reward_mean": 0.921875,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.448275862068965,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.1774193548387097,
      "perf/rollout_time": 241.18918132781982,
      "rollout/response_len/mean": 289.3672456575682,
      "rollout/total_lengths": 8213.300248138958,
      "rollout/raw_reward": 0.6451612903225806,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15669266134500504,
      "rollout/truncated": 0.1774193548387097,
      "rollout/log_probs": -0.15744412690401077,
      "rollout/rewards": -0.16287593872612227,
      "rollout/returns": 4.6566128730773926e-09,
      "rollout/advantages": 4.6566128730773926e-09,
      "rollout/response_lengths": 289.3672456575682,
      "perf/log_probs_time": 131.21214938163757,
      "perf/actor_train_tok_per_s": 13458.915457582367,
      "perf/train_time": 623.9348630905151,
      "perf/actor_train_tflops": 103.66346924658153,
      "perf/wake_up_time": 6.719045639038086,
      "perf/train_wait_time": 263.6429934501648,
      "perf/actor_train_time": 491.8613257408142,
      "perf/update_weights_time": 1.7073471546173096,
      "perf/step_time": 887.5778565406799,
      "perf/sleep_time": 11.182385206222534,
      "perf/data_preprocess_time": 0.12847447395324707,
      "perf/wait_time_ratio": 0.297036470104954,
      "perf/log_probs_tflops": 129.5308187410152
    },
    {
      "step": 28,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 253.92945658640068,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9259259259259259,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.37037037037037035,
      "perf/longest_effective_sample_tokens_per_sec": 16.90750981119609,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/reward_mean": 0.875,
      "rollout/shopsim/success_rate": 0.875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6666666666666666,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 60,
      "rollout/response_len/median": 255,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.3333333333333333,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 28,
      "perf/longest_sample_tokens_per_sec": 16.90750981119609,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "perf/effective_tokens_per_gpu_per_sec": 253.92945658640068,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5925925925925926,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/strict_reward_mean": 0.875,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.037037037037037,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.2551390568319226,
      "perf/rollout_time": 242.25921177864075,
      "rollout/response_len/mean": 297.5417170495768,
      "rollout/total_lengths": 8308.264812575575,
      "rollout/raw_reward": 0.5066505441354293,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15521620213985443,
      "rollout/truncated": 0.2551390568319226,
      "rollout/log_probs": -0.15607056766748428,
      "rollout/rewards": -0.1325012950323513,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/response_lengths": 297.5417170495768,
      "perf/log_probs_time": 137.03238344192505,
      "perf/actor_train_tok_per_s": 13474.944956508705,
      "perf/train_time": 647.7684111595154,
      "perf/actor_train_tflops": 105.3183490381864,
      "perf/wake_up_time": 6.737539768218994,
      "perf/train_wait_time": 264.8340220451355,
      "perf/actor_train_time": 509.90449476242065,
      "perf/update_weights_time": 1.7039895057678223,
      "perf/step_time": 912.6024332046509,
      "perf/sleep_time": 11.458016633987427,
      "perf/data_preprocess_time": 0.12720584869384766,
      "perf/wait_time_ratio": 0.2901964890835948,
      "perf/log_probs_tflops": 130.63165121181734
    },
    {
      "step": 29,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 226.96216108320502,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9090909090909091,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.4090909090909091,
      "perf/longest_effective_sample_tokens_per_sec": 14.7757854576013,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5454545454545454,
      "rollout/shopsim/reward_mean": 0.8390625,
      "rollout/shopsim/success_rate": 0.83203125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6818181818181818,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.45454545454545453,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 63,
      "rollout/response_len/median": 255.5,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.3125,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.22727272727272727,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 29,
      "perf/longest_sample_tokens_per_sec": 14.7757854576013,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "perf/effective_tokens_per_gpu_per_sec": 226.96216108320502,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.8390625,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.727272727272727,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.2523148148148148,
      "perf/rollout_time": 277.2103054523468,
      "rollout/response_len/mean": 291.27893518518516,
      "rollout/total_lengths": 8351.248842592593,
      "rollout/raw_reward": 0.5025462962962963,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15084920078516006,
      "rollout/truncated": 0.2523148148148148,
      "rollout/log_probs": -0.15166538953781128,
      "rollout/rewards": -0.1351905542391318,
      "rollout/returns": -1.4901161193847656e-08,
      "rollout/advantages": -1.4901161193847656e-08,
      "rollout/response_lengths": 291.27893518518516,
      "perf/log_probs_time": 145.93143725395203,
      "perf/actor_train_tok_per_s": 13464.205375226196,
      "perf/train_time": 682.3981907367706,
      "perf/actor_train_tflops": 104.783388349336,
      "perf/wake_up_time": 6.879742622375488,
      "perf/train_wait_time": 299.9490613937378,
      "perf/actor_train_time": 535.9008421897888,
      "perf/update_weights_time": 1.7180087566375732,
      "perf/step_time": 982.3472521305084,
      "perf/sleep_time": 11.146339893341064,
      "perf/data_preprocess_time": 0.13851380348205566,
      "perf/wait_time_ratio": 0.3053391361794011,
      "perf/log_probs_tflops": 128.26458580040733
    },
    {
      "step": 30,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 251.9352882387047,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9565217391304348,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.6086956521739131,
      "perf/longest_effective_sample_tokens_per_sec": 6.331474505076193,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1423,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6956521739130435,
      "rollout/shopsim/reward_mean": 0.8367660984848485,
      "rollout/shopsim/success_rate": 0.83203125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7391304347826086,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.391304347826087,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 65,
      "rollout/response_len/median": 246,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.28125,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.13043478260869565,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 30,
      "perf/longest_sample_tokens_per_sec": 6.331474505076193,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "perf/effective_tokens_per_gpu_per_sec": 251.9352882387047,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.43478260869565216,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/strict_reward_mean": 0.8367660984848485,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.9565217391304346,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.14392059553349876,
      "perf/rollout_time": 224.7501745223999,
      "rollout/response_len/mean": 281.0049627791563,
      "rollout/total_lengths": 7704.419354838709,
      "rollout/raw_reward": 0.5305662079855629,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15510831028223038,
      "rollout/truncated": 0.14392059553349876,
      "rollout/log_probs": -0.1558360457420349,
      "rollout/rewards": -0.11602485764115383,
      "rollout/returns": 3.725290298461914e-09,
      "rollout/advantages": 3.725290298461914e-09,
      "rollout/response_lengths": 281.0049627791563,
      "perf/log_probs_time": 121.19080376625061,
      "perf/actor_train_tok_per_s": 13556.791469381276,
      "perf/train_time": 579.886702299118,
      "perf/actor_train_tflops": 102.13775435864156,
      "perf/wake_up_time": 6.766808271408081,
      "perf/train_wait_time": 247.65843033790588,
      "perf/actor_train_time": 458.05543398857117,
      "perf/update_weights_time": 1.7161118984222412,
      "perf/step_time": 827.5451326370239,
      "perf/sleep_time": 11.461233854293823,
      "perf/data_preprocess_time": 0.1237027645111084,
      "perf/wait_time_ratio": 0.2992687897863974,
      "perf/log_probs_tflops": 128.6807027855643
    },
    {
      "step": 31,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 218.8037229378063,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.84,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.32,
      "perf/longest_effective_sample_tokens_per_sec": 4.215880972057222,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1266,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6,
      "rollout/shopsim/reward_mean": 0.6897631448412699,
      "rollout/shopsim/success_rate": 0.66796875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.52,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 29,
      "rollout/response_len/median": 214,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.44,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 31,
      "perf/longest_sample_tokens_per_sec": 4.215880972057222,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 6,
      "rollout/zero_std/count_0.4": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.25,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.25,
      "perf/effective_tokens_per_gpu_per_sec": 218.8037229378063,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.56,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.25,
      "rollout/zero_std/count_1.0": 2,
      "rollout/shopsim/strict_reward_mean": 0.6897631448412699,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.88,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.2910927456382002,
      "perf/rollout_time": 300.2931079864502,
      "rollout/response_len/mean": 241.34159779614325,
      "rollout/total_lengths": 9492.602387511479,
      "rollout/raw_reward": 0.2852843004358156,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.1641978919506073,
      "rollout/truncated": 0.2910927456382002,
      "rollout/log_probs": -0.16574081778526306,
      "rollout/rewards": -0.0494168752700063,
      "rollout/returns": -0.0018626454984769225,
      "rollout/advantages": -0.0018626454984769225,
      "rollout/response_lengths": 241.34159779614325,
      "perf/log_probs_time": 203.43732357025146,
      "perf/actor_train_tok_per_s": 13424.00482536669,
      "perf/train_time": 974.3193120956421,
      "perf/actor_train_tflops": 107.99431719802173,
      "perf/wake_up_time": 6.794486045837402,
      "perf/train_wait_time": 322.9740581512451,
      "perf/actor_train_time": 770.071534872055,
      "perf/update_weights_time": 1.7132477760314941,
      "perf/step_time": 1297.2933702468872,
      "perf/sleep_time": 11.221829652786255,
      "perf/data_preprocess_time": 0.1738758087158203,
      "perf/wait_time_ratio": 0.24895992345184043,
      "perf/log_probs_tflops": 136.26367102924453
    },
    {
      "step": 32,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 258.3134780684143,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9523809523809523,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.42857142857142855,
      "perf/longest_effective_sample_tokens_per_sec": 16.255403252737203,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5714285714285714,
      "rollout/shopsim/reward_mean": 0.8125,
      "rollout/shopsim/success_rate": 0.8125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6190476190476191,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.47619047619047616,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 37,
      "rollout/response_len/median": 237,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.34375,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23809523809523808,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 32,
      "perf/longest_sample_tokens_per_sec": 16.255403252737203,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 258.3134780684143,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5714285714285714,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.8125,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.857142857142857,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.34291581108829566,
      "perf/rollout_time": 251.97775387763977,
      "rollout/response_len/mean": 267.3069815195072,
      "rollout/total_lengths": 9364.108829568788,
      "rollout/raw_reward": 0.4004106776180698,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.16034730523824692,
      "rollout/truncated": 0.34291581108829566,
      "rollout/log_probs": -0.16198233515024185,
      "rollout/rewards": -0.17251242865527192,
      "rollout/returns": -8.847564458847046e-09,
      "rollout/advantages": -8.847564458847046e-09,
      "rollout/response_lengths": 267.3069815195072,
      "perf/log_probs_time": 181.77904987335205,
      "perf/actor_train_tok_per_s": 13271.82583820836,
      "perf/train_time": 869.6224229335785,
      "perf/actor_train_tflops": 106.74608366056215,
      "perf/wake_up_time": 6.8644373416900635,
      "perf/train_wait_time": 275.16794443130493,
      "perf/actor_train_time": 687.2183308601379,
      "perf/update_weights_time": 1.7318191528320312,
      "perf/step_time": 1144.7903673648834,
      "perf/sleep_time": 11.688857078552246,
      "perf/data_preprocess_time": 0.15407586097717285,
      "perf/wait_time_ratio": 0.24036535620464353,
      "perf/log_probs_tflops": 134.51837178189234
    },
    {
      "step": 33,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 209.54239796025735,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.896551724137931,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5172413793103449,
      "perf/longest_effective_sample_tokens_per_sec": 2.8963202098540575,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 864,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6551724137931034,
      "rollout/shopsim/reward_mean": 0.8302951388888888,
      "rollout/shopsim/success_rate": 0.828125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6896551724137931,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5172413793103449,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 25,
      "rollout/response_len/median": 234,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.09375,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.27586206896551724,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 33,
      "perf/longest_sample_tokens_per_sec": 2.8963202098540575,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 209.54239796025735,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5862068965517241,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 7,
      "rollout/shopsim/strict_reward_mean": 0.8302951388888888,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.137931034482759,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.40398740818467993,
      "perf/rollout_time": 298.30955743789673,
      "rollout/response_len/mean": 262.3651626442812,
      "rollout/total_lengths": 9262.741867785939,
      "rollout/raw_reward": 0.4103999067272939,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15579473227262497,
      "rollout/truncated": 0.40398740818467993,
      "rollout/log_probs": -0.15666131675243378,
      "rollout/rewards": -0.04380570700761529,
      "rollout/returns": 7.450580596923828e-09,
      "rollout/advantages": 7.450580596923828e-09,
      "rollout/response_lengths": 262.3651626442812,
      "perf/log_probs_time": 178.12406039237976,
      "perf/actor_train_tok_per_s": 13354.583211960036,
      "perf/train_time": 839.8394041061401,
      "perf/actor_train_tflops": 107.42531728587474,
      "perf/wake_up_time": 6.858383417129517,
      "perf/train_wait_time": 321.22406339645386,
      "perf/actor_train_time": 661.0010106563568,
      "perf/update_weights_time": 1.7123219966888428,
      "perf/step_time": 1161.063467502594,
      "perf/sleep_time": 11.550317287445068,
      "perf/data_preprocess_time": 0.15838193893432617,
      "perf/wait_time_ratio": 0.2766636556805937,
      "perf/log_probs_tflops": 132.88162407635596
    },
    {
      "step": 34,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 253.78720414122643,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.9230769230769231,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5384615384615384,
      "perf/longest_effective_sample_tokens_per_sec": 5.826707814723956,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 1137,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6923076923076923,
      "rollout/shopsim/reward_mean": 0.7734375,
      "rollout/shopsim/success_rate": 0.76171875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.8076923076923077,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 77,
      "rollout/response_len/median": 235,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.34615384615384615,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 34,
      "perf/longest_sample_tokens_per_sec": 5.826707814723956,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "perf/effective_tokens_per_gpu_per_sec": 253.78720414122643,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.4230769230769231,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.7734375,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.115384615384615,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.08906882591093117,
      "perf/rollout_time": 195.13592171669006,
      "rollout/response_len/mean": 267.3306342780027,
      "rollout/total_lengths": 7055.0661268556005,
      "rollout/raw_reward": 0.47638326585695007,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15886962413787842,
      "rollout/truncated": 0.08906882591093117,
      "rollout/log_probs": -0.15967365354299545,
      "rollout/rewards": -0.10132558653872666,
      "rollout/returns": 1.3969838619232178e-08,
      "rollout/advantages": 1.3969838619232178e-08,
      "rollout/response_lengths": 267.3306342780027,
      "perf/log_probs_time": 105.25784420967102,
      "perf/actor_train_tok_per_s": 13602.351753224499,
      "perf/train_time": 490.3503670692444,
      "perf/actor_train_tflops": 100.17880018382007,
      "perf/wake_up_time": 7.754684686660767,
      "perf/train_wait_time": 219.31555318832397,
      "perf/actor_train_time": 384.33089327812195,
      "perf/update_weights_time": 1.725248098373413,
      "perf/step_time": 709.6659202575684,
      "perf/sleep_time": 11.526955842971802,
      "perf/data_preprocess_time": 0.1074075698852539,
      "perf/wait_time_ratio": 0.30904055968859956,
      "perf/log_probs_tflops": 121.92854620090009
    },
    {
      "step": 35,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 215.27971343781832,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.7857142857142857,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.5714285714285714,
      "perf/longest_effective_sample_tokens_per_sec": 16.581152806342683,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5,
      "rollout/shopsim/reward_mean": 0.875,
      "rollout/shopsim/success_rate": 0.875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7142857142857143,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5714285714285714,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 34,
      "rollout/response_len/median": 229,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2857142857142857,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 35,
      "perf/longest_sample_tokens_per_sec": 16.581152806342683,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 215.27971343781832,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.39285714285714285,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/strict_reward_mean": 0.875,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.8214285714285716,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.32142857142857145,
      "perf/rollout_time": 247.02745628356934,
      "rollout/response_len/mean": 253.23809523809524,
      "rollout/total_lengths": 8638.62619047619,
      "rollout/raw_reward": 0.5095238095238095,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.15085045993328094,
      "rollout/truncated": 0.32142857142857145,
      "rollout/log_probs": -0.152912475168705,
      "rollout/rewards": -0.046068246449743,
      "rollout/returns": 9.313225746154785e-10,
      "rollout/advantages": 9.313225746154785e-10,
      "rollout/response_lengths": 253.23809523809524,
      "perf/log_probs_time": 144.47595191001892,
      "perf/actor_train_tok_per_s": 13509.45006299616,
      "perf/train_time": 682.2091314792633,
      "perf/actor_train_tflops": 107.19717014332352,
      "perf/wake_up_time": 6.432230234146118,
      "perf/train_wait_time": 269.12381649017334,
      "perf/actor_train_time": 537.1385190486908,
      "perf/update_weights_time": 1.7190651893615723,
      "perf/step_time": 951.3329479694366,
      "perf/sleep_time": 11.071479797363281,
      "perf/data_preprocess_time": 0.13090181350708008,
      "perf/wait_time_ratio": 0.2828913022140167,
      "perf/log_probs_tflops": 132.84732039201597
    },
    {
      "step": 36,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 262.9661523045482,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.28,
      "perf/longest_effective_sample_tokens_per_sec": 5.874667406745521,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 923,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.48,
      "rollout/shopsim/reward_mean": 0.8919270833333334,
      "rollout/shopsim/success_rate": 0.88671875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.52,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.56,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 94,
      "rollout/response_len/median": 238,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.52,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 36,
      "perf/longest_sample_tokens_per_sec": 5.874667406745521,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "perf/effective_tokens_per_gpu_per_sec": 262.9661523045482,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.28,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/strict_reward_mean": 0.8919270833333334,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.52,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.06382978723404255,
      "perf/rollout_time": 157.11527752876282,
      "rollout/response_len/mean": 270.48117839607204,
      "rollout/total_lengths": 5861.921440261865,
      "rollout/raw_reward": 0.696126568466994,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.14880386739969254,
      "rollout/truncated": 0.06382978723404255,
      "rollout/log_probs": -0.14932487905025482,
      "rollout/rewards": -0.07928266687205887,
      "rollout/returns": 5.587935447692871e-09,
      "rollout/advantages": 5.587935447692871e-09,
      "rollout/response_lengths": 270.48117839607204,
      "perf/log_probs_time": 70.88788294792175,
      "perf/actor_train_tok_per_s": 13708.053692990945,
      "perf/train_time": 332.9852223396301,
      "perf/actor_train_tflops": 95.49458851228474,
      "perf/wake_up_time": 6.553093194961548,
      "perf/train_wait_time": 178.68143320083618,
      "perf/actor_train_time": 261.27954268455505,
      "perf/update_weights_time": 1.711395263671875,
      "perf/step_time": 511.6666555404663,
      "perf/sleep_time": 10.83518648147583,
      "perf/data_preprocess_time": 0.0838625431060791,
      "perf/wait_time_ratio": 0.3492145350220203,
      "perf/log_probs_tflops": 117.32509317410319
    },
    {
      "step": 37,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 256.682992639953,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8846153846153846,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.46153846153846156,
      "perf/longest_effective_sample_tokens_per_sec": 4.882267282907092,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 939,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.6153846153846154,
      "rollout/shopsim/reward_mean": 0.8015625,
      "rollout/shopsim/success_rate": 0.796875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5384615384615384,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 81,
      "rollout/response_len/median": 246,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.34615384615384615,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 37,
      "perf/longest_sample_tokens_per_sec": 4.882267282907092,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "perf/effective_tokens_per_gpu_per_sec": 256.682992639953,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5769230769230769,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/strict_reward_mean": 0.8015625,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.923076923076923,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.08251748251748252,
      "perf/rollout_time": 192.32867550849915,
      "rollout/response_len/mean": 276.1818181818182,
      "rollout/total_lengths": 7017.823776223776,
      "rollout/raw_reward": 0.5206993006993007,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.14517906308174133,
      "rollout/truncated": 0.08251748251748252,
      "rollout/log_probs": -0.14609608799219131,
      "rollout/rewards": -0.11341218302299927,
      "rollout/returns": 1.3969838619232178e-09,
      "rollout/advantages": 1.3969838619232178e-09,
      "rollout/response_lengths": 276.1818181818182,
      "perf/log_probs_time": 98.62288761138916,
      "perf/actor_train_tok_per_s": 13549.568882111591,
      "perf/train_time": 470.1972894668579,
      "perf/actor_train_tflops": 99.43352004773875,
      "perf/wake_up_time": 6.401993274688721,
      "perf/train_wait_time": 214.37824273109436,
      "perf/actor_train_time": 370.32499289512634,
      "perf/update_weights_time": 1.727567195892334,
      "perf/step_time": 684.5755321979523,
      "perf/sleep_time": 11.126230955123901,
      "perf/data_preprocess_time": 0.10188102722167969,
      "perf/wait_time_ratio": 0.31315498823452637,
      "perf/log_probs_tflops": 124.4562950752752
    },
    {
      "step": 38,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 210.5751395824309,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8461538461538461,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.4230769230769231,
      "perf/longest_effective_sample_tokens_per_sec": 4.464892958473105,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 745,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.7307692307692307,
      "rollout/shopsim/reward_mean": 0.84375,
      "rollout/shopsim/success_rate": 0.828125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5384615384615384,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.4230769230769231,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 73,
      "rollout/response_len/median": 214,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 38,
      "perf/longest_sample_tokens_per_sec": 4.464892958473105,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 210.5751395824309,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.5,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 7,
      "rollout/shopsim/strict_reward_mean": 0.84375,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.6923076923076925,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.075,
      "perf/rollout_time": 166.85730361938477,
      "rollout/response_len/mean": 234.24,
      "rollout/total_lengths": 5992.028333333334,
      "rollout/raw_reward": 0.6383333333333333,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.14057446271181107,
      "rollout/truncated": 0.075,
      "rollout/log_probs": -0.14129453897476196,
      "rollout/rewards": -0.07836518476406733,
      "rollout/returns": -9.313225746154785e-10,
      "rollout/advantages": -9.313225746154785e-10,
      "rollout/response_lengths": 234.24,
      "perf/log_probs_time": 70.87332534790039,
      "perf/actor_train_tok_per_s": 13741.246206800055,
      "perf/train_time": 333.33548164367676,
      "perf/actor_train_tflops": 97.50150495484962,
      "perf/wake_up_time": 6.592628002166748,
      "perf/train_wait_time": 188.65080428123474,
      "perf/actor_train_time": 261.6368956565857,
      "perf/update_weights_time": 1.7156670093536377,
      "perf/step_time": 521.9862859249115,
      "perf/sleep_time": 11.080156564712524,
      "perf/data_preprocess_time": 0.08692312240600586,
      "perf/wait_time_ratio": 0.3614095032151332,
      "perf/log_probs_tflops": 119.97927736097961
    },
    {
      "step": 39,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 194.21942071161058,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8928571428571429,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.2857142857142857,
      "perf/longest_effective_sample_tokens_per_sec": 3.2618549184519403,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 692,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.5357142857142857,
      "rollout/shopsim/reward_mean": 0.845703125,
      "rollout/shopsim/success_rate": 0.828125,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6785714285714286,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.4642857142857143,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 60,
      "rollout/response_len/median": 205,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.5,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 39,
      "perf/longest_sample_tokens_per_sec": 3.2618549184519403,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/zero_std/count_0.5": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "perf/effective_tokens_per_gpu_per_sec": 194.21942071161058,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.6071428571428571,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/strict_reward_mean": 0.845703125,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.9642857142857144,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.24295302013422818,
      "perf/rollout_time": 212.14922714233398,
      "rollout/response_len/mean": 221.2268456375839,
      "rollout/total_lengths": 7569.35033557047,
      "rollout/raw_reward": 0.5671140939597316,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.1479572206735611,
      "rollout/truncated": 0.24295302013422818,
      "rollout/log_probs": -0.14863993972539902,
      "rollout/rewards": -0.06407204362370024,
      "rollout/returns": -2.0954757928848267e-09,
      "rollout/advantages": -2.0954757928848267e-09,
      "rollout/response_lengths": 221.2268456375839,
      "perf/log_probs_time": 113.10249853134155,
      "perf/actor_train_tok_per_s": 13596.368240941682,
      "perf/train_time": 528.7053854465485,
      "perf/actor_train_tflops": 104.14323601880271,
      "perf/wake_up_time": 6.517659902572632,
      "perf/train_wait_time": 234.26168656349182,
      "perf/actor_train_time": 414.75531554222107,
      "perf/update_weights_time": 1.7203609943389893,
      "perf/step_time": 762.9670720100403,
      "perf/sleep_time": 11.34818959236145,
      "perf/data_preprocess_time": 0.10616254806518555,
      "perf/wait_time_ratio": 0.3070403627594679,
      "perf/log_probs_tflops": 127.30034342727083
    },
    {
      "step": 40,
      "round": 1,
      "run_id": "chjizc8u",
      "perf/tokens_per_gpu_per_sec": 216.22349470317388,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.8888888888888888,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/inclusion_rate": 0.48148148148148145,
      "perf/longest_effective_sample_tokens_per_sec": 2.968808279461823,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/response_len/max": 578,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/inclusion_rate": 0.7407407407407407,
      "rollout/shopsim/reward_mean": 0.82734375,
      "rollout/shopsim/success_rate": 0.82421875,
      "rollout/shopsim/chunk/chunk-bbe6cb9408eb/target_rho": 0.589599778982441,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7777777777777778,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6296296296296297,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/rho_min": 0.3445315989317448,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/response_len/min": 62,
      "rollout/response_len/median": 207,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4444444444444444,
      "rollout/shopsim/trajectory_count": 256,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.5449789167055454,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.8999999999999999,
      "rollout/step": 40,
      "perf/longest_sample_tokens_per_sec": 2.968808279461823,
      "rollout/repetition_frac": 0,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6423226971958329,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5190166131192998,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "perf/effective_tokens_per_gpu_per_sec": 216.22349470317388,
      "rollout/shopsim/chunk/chunk-27a7e5bb9344/target_rho": 0.4558757756140158,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/zero_std/count_1.0": 4,
      "rollout/shopsim/strict_reward_mean": 0.82734375,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.444444444444445,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.3445315989317448,
      "rollout/shopsim/rho_mean": 0.570903625792697,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/truncated_ratio": 0.13704206241519673,
      "perf/rollout_time": 194.69091486930847,
      "rollout/response_len/mean": 228.4762550881954,
      "rollout/total_lengths": 7071.377204884668,
      "rollout/raw_reward": 0.6067842605156037,
      "rollout/kl": 0,
      "rollout/rollout_log_probs": -0.1505615934729576,
      "rollout/truncated": 0.13704206241519673,
      "rollout/log_probs": -0.15149366855621338,
      "rollout/rewards": -0.013188100411092895,
      "rollout/returns": 2.3283064365386963e-10,
      "rollout/advantages": 2.3283064365386963e-10,
      "rollout/response_lengths": 228.4762550881954,
      "perf/log_probs_time": 104.22773098945618,
      "perf/actor_train_tok_per_s": 13578.896610124099,
      "perf/train_time": 488.8474724292755,
      "perf/actor_train_tflops": 100.95151143152925,
      "perf/wake_up_time": 6.549328804016113,
      "perf/train_wait_time": 216.70664978027344,
      "perf/actor_train_time": 383.80180287361145,
      "perf/update_weights_time": 1.7349355220794678,
      "perf/step_time": 705.554122209549,
      "perf/sleep_time": 11.223007202148438,
      "perf/data_preprocess_time": 0.10515236854553223,
      "perf/wait_time_ratio": 0.3071439071202985,
      "perf/log_probs_tflops": 123.91255098305352
    },
    {
      "step": 41,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 1.956716996304247,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.52,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.92578125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.92,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 6,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 195.49312169312168,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 12,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 493,
      "rollout/shopsim/reward_mean": 0.92578125,
      "rollout/shopsim/strict_reward_mean": 0.92578125,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 251.9526333808899,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 1.956716996304247,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.48,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.37566137566137564,
      "perf/tokens_per_gpu_per_sec": 183.30925690377427,
      "rollout/step": 41,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 183.30925690377427,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 185,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.56,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.24,
      "rollout/rewards": -0.0017144779679636477,
      "rollout/response_lengths": 195.49312169312168,
      "rollout/log_probs": -0.13969792425632477,
      "rollout/returns": 1.1641532182693481e-09,
      "rollout/raw_reward": 0.5693121693121693,
      "rollout/truncated": 0.37566137566137564,
      "rollout/total_lengths": 8574.716402116403,
      "rollout/rollout_log_probs": -0.1388842985033989,
      "rollout/advantages": 1.1641532182693481e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 7.1701436042785645,
      "perf/train_time": 769.1901543140411,
      "perf/actor_train_tok_per_s": 13439.34536302598,
      "perf/update_weights_time": 1.7416954040527344,
      "perf/wait_time_ratio": 0.25208081326352527,
      "perf/step_time": 1028.4401950836182,
      "perf/train_wait_time": 259.250040769577,
      "perf/wake_up_time": 4.1757118701934814,
      "perf/data_preprocess_time": 0.1476764678955078,
      "perf/actor_train_time": 602.9391150474548,
      "perf/log_probs_tflops": 130.39273569781423,
      "perf/log_probs_time": 165.42254495620728,
      "perf/actor_train_tflops": 107.32376277115569
    },
    {
      "step": 42,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 16.39491118351395,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.75,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2857142857142857,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8214285714285714,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 47,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 236.59295499021528,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.76171875,
      "rollout/shopsim/strict_reward_mean": 0.76171875,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 249.8336193561554,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 16.39491118351395,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.6071428571428571,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.44129158512720157,
      "perf/tokens_per_gpu_per_sec": 241.95902919624675,
      "rollout/step": 42,
      "rollout/zero_std/count_0.0": 5,
      "perf/effective_tokens_per_gpu_per_sec": 241.95902919624675,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 206,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.7142857142857144,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5,
      "rollout/rewards": -0.1546564428773645,
      "rollout/response_lengths": 236.59295499021528,
      "rollout/log_probs": -0.14970871061086655,
      "rollout/returns": 0,
      "rollout/raw_reward": 0.3336594911937378,
      "rollout/truncated": 0.44129158512720157,
      "rollout/total_lengths": 9030.76614481409,
      "rollout/rollout_log_probs": -0.14884991943836212,
      "rollout/advantages": 0,
      "rollout/kl": 0,
      "perf/sleep_time": 11.25025486946106,
      "perf/train_time": 869.2017698287964,
      "perf/actor_train_tok_per_s": 13436.318425221341,
      "perf/update_weights_time": 1.7206060886383057,
      "perf/wait_time_ratio": 0.23976064995351398,
      "perf/step_time": 1143.3264665603638,
      "perf/train_wait_time": 274.1246967315674,
      "perf/wake_up_time": 6.060720443725586,
      "perf/data_preprocess_time": 0.15123510360717773,
      "perf/actor_train_time": 686.9026699066162,
      "perf/log_probs_tflops": 135.08526860200232,
      "perf/log_probs_time": 181.56878471374512,
      "perf/actor_train_tflops": 107.12115032016044
    },
    {
      "step": 43,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 15.449212496205247,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.64,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.8203125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.24,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.84,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 67,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 233.68004459308807,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.8296875,
      "rollout/shopsim/strict_reward_mean": 0.8296875,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 265.12678241729736,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 15.449212496205247,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.68,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.12709030100334448,
      "perf/tokens_per_gpu_per_sec": 197.65166507214832,
      "rollout/step": 43,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 197.65166507214832,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 199,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.88,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.48,
      "rollout/rewards": 0.0060873646019161615,
      "rollout/response_lengths": 233.68004459308807,
      "rollout/log_probs": -0.15264662355184555,
      "rollout/returns": -1.6763806343078613e-08,
      "rollout/raw_reward": 0.576662950575994,
      "rollout/truncated": 0.12709030100334448,
      "rollout/total_lengths": 7877.108138238573,
      "rollout/rollout_log_probs": -0.15196417272090912,
      "rollout/advantages": -1.6763806343078613e-08,
      "rollout/kl": 0,
      "perf/sleep_time": 10.689144849777222,
      "perf/train_time": 675.3845107555389,
      "perf/actor_train_tok_per_s": 13224.582909931212,
      "perf/update_weights_time": 1.7258405685424805,
      "perf/wait_time_ratio": 0.29796643519285776,
      "perf/step_time": 962.0401995182037,
      "perf/train_wait_time": 286.6556887626648,
      "perf/wake_up_time": 6.042240142822266,
      "perf/data_preprocess_time": 0.1368870735168457,
      "perf/actor_train_time": 534.2902719974518,
      "perf/log_probs_tflops": 128.11749365081585,
      "perf/log_probs_time": 140.30880618095398,
      "perf/actor_train_tflops": 100.93396900811057
    },
    {
      "step": 44,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 19.625059833184945,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.81640625,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.34615384615384615,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8076923076923077,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 57,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 212.75239398084815,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.818359375,
      "rollout/shopsim/strict_reward_mean": 0.818359375,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 208.71273946762085,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 19.625059833184945,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.6538461538461539,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.12448700410396717,
      "perf/tokens_per_gpu_per_sec": 186.28714327249688,
      "rollout/step": 44,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 186.28714327249688,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 179,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.423076923076923,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.38461538461538464,
      "rollout/rewards": -0.09987912519685874,
      "rollout/response_lengths": 212.75239398084815,
      "rollout/log_probs": -0.1437220424413681,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/raw_reward": 0.5191518467852257,
      "rollout/truncated": 0.12448700410396717,
      "rollout/total_lengths": 6571.6456908344735,
      "rollout/rollout_log_probs": -0.14302929490804672,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.877238512039185,
      "perf/train_time": 449.9418535232544,
      "perf/actor_train_tok_per_s": 13573.51136763834,
      "perf/update_weights_time": 1.717872142791748,
      "perf/wait_time_ratio": 0.3384318398624368,
      "perf/step_time": 680.114129781723,
      "perf/train_wait_time": 230.17227625846863,
      "perf/wake_up_time": 6.1342785358428955,
      "perf/data_preprocess_time": 0.10979175567626953,
      "perf/actor_train_time": 353.9152743816376,
      "perf/log_probs_tflops": 122.48597343364327,
      "perf/log_probs_time": 95.18854856491089,
      "perf/actor_train_tflops": 98.8309593397447
    },
    {
      "step": 45,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.503438096732676,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.52,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.90625,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.92,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 63,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 197.83916083916083,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 619,
      "rollout/shopsim/reward_mean": 0.91748046875,
      "rollout/shopsim/strict_reward_mean": 0.91748046875,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 176.68358421325684,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.503438096732676,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.76,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.08951048951048951,
      "perf/tokens_per_gpu_per_sec": 200.1530032202426,
      "rollout/step": 45,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 200.1530032202426,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 179,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 3,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.4,
      "rollout/rewards": -0.010891853095768214,
      "rollout/response_lengths": 197.83916083916083,
      "rollout/log_probs": -0.14385036379098892,
      "rollout/returns": 0,
      "rollout/raw_reward": 0.7013986013986014,
      "rollout/truncated": 0.08951048951048951,
      "rollout/total_lengths": 6553.071328671329,
      "rollout/rollout_log_probs": -0.14307260513305664,
      "rollout/advantages": 0,
      "rollout/kl": 0,
      "perf/sleep_time": 11.063018083572388,
      "perf/train_time": 438.73643803596497,
      "perf/actor_train_tok_per_s": 13609.506726064223,
      "perf/update_weights_time": 1.7089900970458984,
      "perf/wait_time_ratio": 0.31257324161522765,
      "perf/step_time": 638.2300844192505,
      "perf/train_wait_time": 199.49364638328552,
      "perf/wake_up_time": 6.048165321350098,
      "perf/data_preprocess_time": 0.10983157157897949,
      "perf/actor_train_time": 344.27743005752563,
      "perf/log_probs_tflops": 121.73285673133161,
      "perf/log_probs_time": 93.60923647880554,
      "perf/actor_train_tflops": 99.29770683282712
    },
    {
      "step": 46,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.6286599626346563,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5454545454545454,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.81640625,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4090909090909091,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9545454545454546,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 48,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 206.47283126787417,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 4,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 1024,
      "rollout/shopsim/reward_mean": 0.8232421875,
      "rollout/shopsim/strict_reward_mean": 0.8232421875,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 282.19783902168274,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.6286599626346563,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.6818181818181818,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.223069590085796,
      "perf/tokens_per_gpu_per_sec": 191.8777981706641,
      "rollout/step": 46,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 191.8777981706641,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.3125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 185,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.9545454545454546,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.36363636363636365,
      "rollout/rewards": -0.08590311908631011,
      "rollout/response_lengths": 206.47283126787417,
      "rollout/log_probs": -0.14450087398290634,
      "rollout/returns": 2.3283064365386963e-09,
      "rollout/raw_reward": 0.5142993326978075,
      "rollout/truncated": 0.223069590085796,
      "rollout/total_lengths": 8420.083889418494,
      "rollout/rollout_log_probs": -0.1436624675989151,
      "rollout/advantages": 2.3283064365386963e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 11.04187536239624,
      "perf/train_time": 834.9753975868225,
      "perf/actor_train_tok_per_s": 13388.669601248135,
      "perf/update_weights_time": 1.7128236293792725,
      "perf/wait_time_ratio": 0.26907414710656097,
      "perf/step_time": 1142.3530776500702,
      "perf/train_wait_time": 307.3776800632477,
      "perf/wake_up_time": 8.150717735290527,
      "perf/data_preprocess_time": 0.16419577598571777,
      "perf/actor_train_time": 659.7121493816376,
      "perf/log_probs_tflops": 131.2554581849032,
      "perf/log_probs_time": 173.28442668914795,
      "perf/actor_train_tflops": 103.4293222098971
    },
    {
      "step": 47,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 9.240376449444224,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.9296875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.23333333333333334,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 60,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 180.01941747572815,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 11,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 1312,
      "rollout/shopsim/reward_mean": 0.948828125,
      "rollout/zero_std/count_0.5": 1,
      "rollout/shopsim/strict_reward_mean": 0.948828125,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 141.98555731773376,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 9.240376449444224,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.7333333333333333,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0,
      "perf/tokens_per_gpu_per_sec": 195.88612057042087,
      "rollout/step": 47,
      "perf/effective_tokens_per_gpu_per_sec": 195.88612057042087,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 157.5,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.8,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.43333333333333335,
      "rollout/rewards": 0.004689754309391898,
      "rollout/response_lengths": 180.01941747572815,
      "rollout/log_probs": -0.13025563210248947,
      "rollout/returns": 0,
      "rollout/raw_reward": 0.906957928802589,
      "rollout/truncated": 0,
      "rollout/total_lengths": 4990.775080906149,
      "rollout/rollout_log_probs": -0.1297425627708435,
      "rollout/advantages": 0,
      "rollout/kl": 0,
      "perf/sleep_time": 11.261091470718384,
      "perf/train_time": 279.29860615730286,
      "perf/actor_train_tok_per_s": 13959.616697463496,
      "perf/update_weights_time": 1.7101242542266846,
      "perf/wait_time_ratio": 0.371974707693897,
      "perf/step_time": 444.72509241104126,
      "perf/train_wait_time": 165.4264862537384,
      "perf/wake_up_time": 7.780729532241821,
      "perf/data_preprocess_time": 0.08464980125427246,
      "perf/actor_train_time": 220.94439029693604,
      "perf/log_probs_tflops": 117.62339670968495,
      "perf/log_probs_time": 57.615601539611816,
      "perf/actor_train_tflops": 92.01785228563291
    },
    {
      "step": 48,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 17.48872133186353,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.7083333333333334,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.796875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4166666666666667,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9166666666666666,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 59,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 232.3417142857143,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.8125,
      "rollout/zero_std/count_0.5": 1,
      "rollout/shopsim/strict_reward_mean": 0.8125,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 234.2080888748169,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 17.48872133186353,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.625,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.43314285714285716,
      "perf/tokens_per_gpu_per_sec": 217.00680896280053,
      "rollout/step": 48,
      "rollout/zero_std/count_0.0": 4,
      "perf/effective_tokens_per_gpu_per_sec": 217.00680896280053,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 179,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.9583333333333335,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.2916666666666667,
      "rollout/rewards": -0.15475492535318647,
      "rollout/response_lengths": 232.3417142857143,
      "rollout/log_probs": -0.1280287206172943,
      "rollout/returns": 3.725290298461914e-09,
      "rollout/raw_reward": 0.38171428571428573,
      "rollout/truncated": 0.43314285714285716,
      "rollout/total_lengths": 8693.57142857143,
      "rollout/rollout_log_probs": -0.12737714126706123,
      "rollout/advantages": 3.725290298461914e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.803268432617188,
      "perf/train_time": 721.6889276504517,
      "perf/actor_train_tok_per_s": 13378.001729626718,
      "perf/update_weights_time": 1.7154622077941895,
      "perf/wait_time_ratio": 0.2635110981710547,
      "perf/step_time": 979.9046881198883,
      "perf/train_wait_time": 258.21576046943665,
      "perf/wake_up_time": 8.171801567077637,
      "perf/data_preprocess_time": 0.1373579502105713,
      "perf/actor_train_time": 568.6107053756714,
      "perf/log_probs_tflops": 134.30820947499518,
      "perf/log_probs_time": 151.1484670639038,
      "perf/actor_train_tflops": 107.10568646169449
    },
    {
      "step": 49,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 15.72413802975689,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.78125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2916666666666667,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9166666666666666,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 55,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 222.74425574425575,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.7972005208333334,
      "rollout/shopsim/strict_reward_mean": 0.7972005208333334,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 260.49122643470764,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 15.72413802975689,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.625,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.20679320679320679,
      "perf/tokens_per_gpu_per_sec": 213.9870534717288,
      "rollout/step": 49,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 213.9870534717288,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 178,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.5833333333333335,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.25,
      "rollout/rewards": -0.13498844833045334,
      "rollout/response_lengths": 222.74425574425575,
      "rollout/log_probs": -0.14605706185102463,
      "rollout/returns": -5.820766091346741e-09,
      "rollout/raw_reward": 0.418997668997669,
      "rollout/truncated": 0.20679320679320679,
      "rollout/total_lengths": 8574.606393606393,
      "rollout/rollout_log_probs": -0.1452210396528244,
      "rollout/advantages": -5.820766091346741e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.736064195632935,
      "perf/train_time": 822.6050560474396,
      "perf/actor_train_tok_per_s": 13200.124318401879,
      "perf/update_weights_time": 1.7110190391540527,
      "perf/wait_time_ratio": 0.25571608817932523,
      "perf/step_time": 1105.230199098587,
      "perf/train_wait_time": 282.62514305114746,
      "perf/wake_up_time": 6.0848164558410645,
      "perf/data_preprocess_time": 0.15302109718322754,
      "perf/actor_train_time": 650.2348608970642,
      "perf/log_probs_tflops": 131.3852674382904,
      "perf/log_probs_time": 170.8489372730255,
      "perf/actor_train_tflops": 103.56427191951641
    },
    {
      "step": 50,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 20.57880187710048,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5357142857142857,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.83203125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.32142857142857145,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9285714285714286,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 3,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 190.28878281622912,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.83203125,
      "rollout/shopsim/strict_reward_mean": 0.83203125,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 199.03977036476135,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 20.57880187710048,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.75,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.3233890214797136,
      "perf/tokens_per_gpu_per_sec": 200.28911773231184,
      "rollout/step": 50,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 200.28911773231184,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 156.5,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.0357142857142856,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5,
      "rollout/rewards": 0.0059876356958489425,
      "rollout/response_lengths": 190.28878281622912,
      "rollout/log_probs": -0.13748693466186523,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/raw_reward": 0.4809069212410501,
      "rollout/truncated": 0.3233890214797136,
      "rollout/total_lengths": 8446.850835322195,
      "rollout/rollout_log_probs": -0.1366030052304268,
      "rollout/advantages": 1.862645149230957e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 11.2706880569458,
      "perf/train_time": 668.4548621177673,
      "perf/actor_train_tok_per_s": 13396.540648274527,
      "perf/update_weights_time": 1.7120013236999512,
      "perf/wait_time_ratio": 0.24864863961971453,
      "perf/step_time": 889.6701295375824,
      "perf/train_wait_time": 221.21526741981506,
      "perf/wake_up_time": 6.090229511260986,
      "perf/data_preprocess_time": 0.13426756858825684,
      "perf/actor_train_time": 528.3797650337219,
      "perf/log_probs_tflops": 133.58789125974812,
      "perf/log_probs_time": 139.2266755104065,
      "perf/actor_train_tflops": 105.60017180457358
    },
    {
      "step": 51,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.2755465209418126,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6296296296296297,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.8203125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4074074074074074,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.7777777777777778,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 56,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 204.18470588235294,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 760,
      "rollout/shopsim/reward_mean": 0.8260091145833334,
      "rollout/shopsim/strict_reward_mean": 0.8260091145833334,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 232.02234959602356,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.2755465209418126,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.5185185185185185,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.23058823529411765,
      "perf/tokens_per_gpu_per_sec": 187.004614320756,
      "rollout/step": 51,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 187.004614320756,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 178,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.6296296296296298,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.2962962962962963,
      "rollout/rewards": -0.09558110044283026,
      "rollout/response_lengths": 204.18470588235294,
      "rollout/log_probs": -0.14583663642406464,
      "rollout/returns": -1.4319084584712982e-08,
      "rollout/raw_reward": 0.5268627450980392,
      "rollout/truncated": 0.23058823529411765,
      "rollout/total_lengths": 7806.22705882353,
      "rollout/rollout_log_probs": -0.14501356333494186,
      "rollout/advantages": -1.4319084584712982e-08,
      "rollout/kl": 0,
      "perf/sleep_time": 10.417779684066772,
      "perf/train_time": 625.1816637516022,
      "perf/actor_train_tok_per_s": 13482.080336540783,
      "perf/update_weights_time": 1.702355146408081,
      "perf/wait_time_ratio": 0.28850956104573355,
      "perf/step_time": 878.6929936408997,
      "perf/train_wait_time": 253.51132988929749,
      "perf/wake_up_time": 6.20528769493103,
      "perf/data_preprocess_time": 0.12465667724609375,
      "perf/actor_train_time": 492.15646505355835,
      "perf/log_probs_tflops": 128.84346048326245,
      "perf/log_probs_time": 132.22329354286194,
      "perf/actor_train_tflops": 103.84567453382499
    },
    {
      "step": 52,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.839551983516702,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5769230769230769,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.8046875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9615384615384616,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 61,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 207.24092741935485,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 946,
      "rollout/shopsim/reward_mean": 0.8046875,
      "rollout/shopsim/strict_reward_mean": 0.8046875,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 246.38291239738464,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.839551983516702,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.4230769230769231,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.22379032258064516,
      "perf/tokens_per_gpu_per_sec": 208.60111401356082,
      "rollout/step": 52,
      "rollout/zero_std/count_0.0": 3,
      "perf/effective_tokens_per_gpu_per_sec": 208.60111401356082,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 174,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.8846153846153846,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5384615384615384,
      "rollout/rewards": -0.11736915347677085,
      "rollout/response_lengths": 207.24092741935485,
      "rollout/log_probs": -0.14877718687057495,
      "rollout/returns": -2.9103830456733704e-09,
      "rollout/raw_reward": 0.4949596774193548,
      "rollout/truncated": 0.22379032258064516,
      "rollout/total_lengths": 8520.767137096775,
      "rollout/rollout_log_probs": -0.14789343625307083,
      "rollout/advantages": -2.9103830456733704e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.89197325706482,
      "perf/train_time": 807.6005074977875,
      "perf/actor_train_tok_per_s": 13234.489400996557,
      "perf/update_weights_time": 1.6989872455596924,
      "perf/wait_time_ratio": 0.2499056745859349,
      "perf/step_time": 1076.6652674674988,
      "perf/train_wait_time": 269.0647599697113,
      "perf/wake_up_time": 6.484991550445557,
      "perf/data_preprocess_time": 0.15774130821228027,
      "perf/actor_train_time": 638.6797966957092,
      "perf/log_probs_tflops": 131.1478771071219,
      "perf/log_probs_time": 168.09191799163818,
      "perf/actor_train_tflops": 103.54906316523481
    },
    {
      "step": 53,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 16.40200073968622,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.44,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.81640625,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 56,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 196.41691542288558,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/reward_mean": 0.8546781994047619,
      "rollout/shopsim/strict_reward_mean": 0.8546781994047619,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 249.72563195228577,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 16.40200073968622,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.56,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.19701492537313434,
      "perf/tokens_per_gpu_per_sec": 197.61587793050052,
      "rollout/step": 53,
      "rollout/zero_std/count_0.0": 1,
      "perf/effective_tokens_per_gpu_per_sec": 197.61587793050052,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 170,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.88,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6,
      "rollout/rewards": -0.14810055233352812,
      "rollout/response_lengths": 196.41691542288558,
      "rollout/log_probs": -0.1422938033938408,
      "rollout/returns": 1.3969838619232178e-08,
      "rollout/raw_reward": 0.5826344468135513,
      "rollout/truncated": 0.19701492537313434,
      "rollout/total_lengths": 8597.72736318408,
      "rollout/rollout_log_probs": -0.14126396924257278,
      "rollout/advantages": 1.3969838619232178e-08,
      "rollout/kl": 0,
      "perf/sleep_time": 11.027637720108032,
      "perf/train_time": 818.736891746521,
      "perf/actor_train_tok_per_s": 13366.186584967756,
      "perf/update_weights_time": 1.7088041305541992,
      "perf/wait_time_ratio": 0.24916733260531088,
      "perf/step_time": 1090.4385588169098,
      "perf/train_wait_time": 271.7016670703888,
      "perf/wake_up_time": 6.077710390090942,
      "perf/data_preprocess_time": 0.1502690315246582,
      "perf/actor_train_time": 646.4608244895935,
      "perf/log_probs_tflops": 131.98241757003086,
      "perf/log_probs_time": 171.07011771202087,
      "perf/actor_train_tflops": 104.77780023614902
    },
    {
      "step": 54,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 5.774555512080076,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.56,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.89453125,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.36,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 50,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 209.36864864864864,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 9,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 1174,
      "rollout/shopsim/reward_mean": 0.9045758928571429,
      "rollout/shopsim/strict_reward_mean": 0.9045758928571429,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 203.30569124221802,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 5.774555512080076,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.68,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.15783783783783784,
      "perf/tokens_per_gpu_per_sec": 238.14630915726147,
      "rollout/step": 54,
      "rollout/zero_std/count_0.0": 1,
      "perf/effective_tokens_per_gpu_per_sec": 238.14630915726147,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 173,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.88,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.4,
      "rollout/rewards": -0.023814818826881613,
      "rollout/response_lengths": 209.36864864864864,
      "rollout/log_probs": -0.14241571724414825,
      "rollout/returns": -8.381903171539307e-09,
      "rollout/raw_reward": 0.6781467181467182,
      "rollout/truncated": 0.15783783783783784,
      "rollout/total_lengths": 8065.991351351351,
      "rollout/rollout_log_probs": -0.14145008474588394,
      "rollout/advantages": -8.381903171539307e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.601900815963745,
      "perf/train_time": 700.9787557125092,
      "perf/actor_train_tok_per_s": 13515.21488067171,
      "perf/update_weights_time": 1.7334215641021729,
      "perf/wait_time_ratio": 0.24288849271451837,
      "perf/step_time": 925.8593337535858,
      "perf/train_wait_time": 224.88057804107666,
      "perf/wake_up_time": 6.068498849868774,
      "perf/data_preprocess_time": 0.13901090621948242,
      "perf/actor_train_time": 552.047604560852,
      "perf/log_probs_tflops": 129.4966889409421,
      "perf/log_probs_time": 147.9443256855011,
      "perf/actor_train_tflops": 104.11221875935676
    },
    {
      "step": 55,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0.03125,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 5.854126251078053,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.76171875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.48148148148148145,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8888888888888888,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 53,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 198.93655828707375,
      "rollout/shopsim/full_skill_retry_success_rate": 0.2,
      "rollout/zero_std/count_1.0": 3,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 1863,
      "rollout/shopsim/reward_mean": 0.7652698863636364,
      "rollout/shopsim/strict_reward_mean": 0.7652698863636364,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 318.23707246780396,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 5.854126251078053,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.7037037037037037,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.4329896907216495,
      "perf/tokens_per_gpu_per_sec": 197.0692776729991,
      "rollout/step": 55,
      "rollout/zero_std/count_0.0": 4,
      "perf/effective_tokens_per_gpu_per_sec": 197.0692776729991,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 175,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.925925925925926,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.37037037037037035,
      "rollout/rewards": -0.14969814629615252,
      "rollout/response_lengths": 198.93655828707375,
      "rollout/log_probs": -0.15304439514875412,
      "rollout/returns": -4.889443516731262e-09,
      "rollout/raw_reward": 0.33292480715161127,
      "rollout/truncated": 0.4329896907216495,
      "rollout/total_lengths": 10297.555908009515,
      "rollout/rollout_log_probs": -0.15231778472661972,
      "rollout/advantages": -4.889443516731262e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 11.19690728187561,
      "perf/train_time": 1248.036847114563,
      "perf/actor_train_tok_per_s": 13161.809954681665,
      "perf/update_weights_time": 1.691840410232544,
      "perf/wait_time_ratio": 0.21437801821599606,
      "perf/step_time": 1588.5971574783325,
      "perf/train_wait_time": 340.56031036376953,
      "perf/wake_up_time": 6.089039325714111,
      "perf/data_preprocess_time": 0.20339441299438477,
      "perf/actor_train_time": 986.5830037593842,
      "perf/log_probs_tflops": 137.754714523673,
      "perf/log_probs_time": 257.2120180130005,
      "perf/actor_train_tflops": 107.74207941478039
    },
    {
      "step": 56,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0.09375,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.0960068660908764,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5217391304347826,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.84765625,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2608695652173913,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8260869565217391,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 66,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 186.4501607717042,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 658,
      "rollout/shopsim/reward_mean": 0.8483072916666666,
      "rollout/shopsim/strict_reward_mean": 0.8483072916666666,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 212.5318284034729,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.0960068660908764,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.5217391304347826,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.21864951768488747,
      "perf/tokens_per_gpu_per_sec": 204.62582158489235,
      "rollout/step": 56,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 204.62582158489235,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.28125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 160,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.6956521739130435,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5652173913043478,
      "rollout/rewards": -0.18697517873901007,
      "rollout/response_lengths": 186.4501607717042,
      "rollout/log_probs": -0.15301518887281418,
      "rollout/returns": -1.862645149230957e-09,
      "rollout/raw_reward": 0.5848517327617005,
      "rollout/truncated": 0.21864951768488747,
      "rollout/total_lengths": 8233.228295819936,
      "rollout/rollout_log_probs": -0.1522940993309021,
      "rollout/advantages": -1.862645149230957e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 11.279544353485107,
      "perf/train_time": 727.2757904529572,
      "perf/actor_train_tok_per_s": 13353.243787157828,
      "perf/update_weights_time": 1.7198741436004639,
      "perf/wait_time_ratio": 0.2439197849068389,
      "perf/step_time": 961.9029514789581,
      "perf/train_wait_time": 234.62716102600098,
      "perf/wake_up_time": 6.042749881744385,
      "perf/data_preprocess_time": 0.14020347595214844,
      "perf/actor_train_time": 575.2611217498779,
      "perf/log_probs_tflops": 131.0286604842073,
      "perf/log_probs_time": 151.3580310344696,
      "perf/actor_train_tflops": 103.42558871167714
    },
    {
      "step": 57,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 5.82348137478921,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.375,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.83984375,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.5,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9166666666666666,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 68,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 186.5295081967213,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 801,
      "rollout/shopsim/reward_mean": 0.83984375,
      "rollout/shopsim/strict_reward_mean": 0.83984375,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 137.54658913612366,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 5.82348137478921,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.5833333333333334,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.10655737704918032,
      "perf/tokens_per_gpu_per_sec": 206.80810900987535,
      "rollout/step": 57,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 206.80810900987535,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 166,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.75,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.375,
      "rollout/rewards": -0.02714117726341623,
      "rollout/response_lengths": 186.5295081967213,
      "rollout/log_probs": -0.13604915142059326,
      "rollout/returns": -1.862645149230957e-09,
      "rollout/raw_reward": 0.6491803278688525,
      "rollout/truncated": 0.10655737704918032,
      "rollout/total_lengths": 5879.0311475409835,
      "rollout/rollout_log_probs": -0.13513126969337463,
      "rollout/advantages": -1.862645149230957e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.942262411117554,
      "perf/train_time": 333.4993073940277,
      "perf/actor_train_tok_per_s": 13656.95252117587,
      "perf/update_weights_time": 1.7357404232025146,
      "perf/wait_time_ratio": 0.3229611381650153,
      "perf/step_time": 492.5851767063141,
      "perf/train_wait_time": 159.08586931228638,
      "perf/wake_up_time": 6.056350469589233,
      "perf/data_preprocess_time": 0.09041357040405273,
      "perf/actor_train_time": 262.5921847820282,
      "perf/log_probs_tflops": 121.79403372123014,
      "perf/log_probs_time": 70.13574743270874,
      "perf/actor_train_tflops": 97.58990651195835
    },
    {
      "step": 58,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 5.764171510980675,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.42857142857142855,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.8984375,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.5714285714285714,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9523809523809523,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 38,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 204.71808510638297,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 1020,
      "rollout/shopsim/reward_mean": 0.8984375,
      "rollout/shopsim/strict_reward_mean": 0.8984375,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 176.95517873764038,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 5.764171510980675,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.6666666666666666,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.030585106382978722,
      "perf/tokens_per_gpu_per_sec": 217.4957538657973,
      "rollout/step": 58,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 217.4957538657973,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.34375,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 177,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.0476190476190474,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.42857142857142855,
      "rollout/rewards": -0.05619411218039533,
      "rollout/response_lengths": 204.71808510638297,
      "rollout/log_probs": -0.1426452100276947,
      "rollout/returns": -1.1641532182693481e-09,
      "rollout/raw_reward": 0.6928191489361702,
      "rollout/truncated": 0.030585106382978722,
      "rollout/total_lengths": 6659.815159574468,
      "rollout/rollout_log_probs": -0.14188400655984879,
      "rollout/advantages": -1.1641532182693481e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 10.791692018508911,
      "perf/train_time": 464.0732264518738,
      "perf/actor_train_tok_per_s": 13795.92366686361,
      "perf/update_weights_time": 1.7168073654174805,
      "perf/wait_time_ratio": 0.299583891705204,
      "perf/step_time": 662.5678949356079,
      "perf/train_wait_time": 198.49466848373413,
      "perf/wake_up_time": 5.986079454421997,
      "perf/data_preprocess_time": 0.10790038108825684,
      "perf/actor_train_time": 363.01889753341675,
      "perf/log_probs_tflops": 120.17406538687355,
      "perf/log_probs_time": 100.29953980445862,
      "perf/actor_train_tflops": 99.60971897027586
    },
    {
      "step": 59,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.162886573400153,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.391304347826087,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.8671875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.2608695652173913,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9130434782608695,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 32,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 184.0847240051348,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 715,
      "rollout/shopsim/reward_mean": 0.8683035714285714,
      "rollout/shopsim/strict_reward_mean": 0.8683035714285714,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 226.0593237876892,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.162886573400153,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.782608695652174,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.27471116816431324,
      "perf/tokens_per_gpu_per_sec": 158.58890223731774,
      "rollout/step": 59,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 158.58890223731774,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.28125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 159,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.9130434782608696,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.5652173913043478,
      "rollout/rewards": -0.025062045290473185,
      "rollout/response_lengths": 184.0847240051348,
      "rollout/log_probs": -0.14318972826004028,
      "rollout/returns": 1.1175870895385742e-08,
      "rollout/raw_reward": 0.5778470566660554,
      "rollout/truncated": 0.27471116816431324,
      "rollout/total_lengths": 7389.460847240051,
      "rollout/rollout_log_probs": -0.1425350233912468,
      "rollout/advantages": 1.1175870895385742e-08,
      "rollout/kl": 0,
      "perf/sleep_time": 10.613769054412842,
      "perf/train_time": 540.4341125488281,
      "perf/actor_train_tok_per_s": 13592.623910771292,
      "perf/update_weights_time": 1.7507405281066895,
      "perf/wait_time_ratio": 0.31412814835899333,
      "perf/step_time": 787.9520223140717,
      "perf/train_wait_time": 247.51790976524353,
      "perf/wake_up_time": 6.0597007274627686,
      "perf/data_preprocess_time": 0.11525082588195801,
      "perf/actor_train_time": 423.49365639686584,
      "perf/log_probs_tflops": 126.79942985148784,
      "perf/log_probs_time": 116.10814833641052,
      "perf/actor_train_tflops": 104.29280430381758
    },
    {
      "step": 60,
      "round": 2,
      "run_id": "7vhbv5j0",
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "rollout/shopsim/rho_min": 0.36213934903190265,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.406239919602322,
      "perf/longest_sample_tokens_per_sec": 3.5167599313522873,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.36,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/success_rate": 0.796875,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.36,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/online-bbef274e1764d953/target_rho": 0.5830719338167873,
      "rollout/response_len/min": 62,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/response_len/mean": 208.02853260869566,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/group_count": 32,
      "rollout/response_len/max": 724,
      "rollout/shopsim/reward_mean": 0.8460286458333334,
      "rollout/zero_std/count_0.5": 2,
      "rollout/shopsim/strict_reward_mean": 0.8460286458333334,
      "rollout/shopsim/train_trajectory_count": 128,
      "perf/rollout_time": 205.8713173866272,
      "rollout/shopsim/trajectory_count": 256,
      "perf/longest_effective_sample_tokens_per_sec": 3.5167599313522873,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.5112815560620121,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-bbef274e1764d953/inclusion_rate": 0.6,
      "rollout/shopsim/rho_mean": 0.5525465517026048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/truncated_ratio": 0.24048913043478262,
      "perf/tokens_per_gpu_per_sec": 185.92803740656677,
      "rollout/step": 60,
      "rollout/zero_std/count_0.0": 2,
      "perf/effective_tokens_per_gpu_per_sec": 185.92803740656677,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/train_group_count": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/response_len/median": 185,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.36213934903190265,
      "rollout/shopsim/skills_per_assisted_group_mean": 2.44,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.24,
      "rollout/rewards": -0.030559935024014846,
      "rollout/response_lengths": 208.02853260869566,
      "rollout/log_probs": -0.1452033817768097,
      "rollout/returns": 9.313225746154785e-09,
      "rollout/raw_reward": 0.5242300724637681,
      "rollout/truncated": 0.24048913043478262,
      "rollout/total_lengths": 7273.077445652174,
      "rollout/rollout_log_probs": -0.1445097252726555,
      "rollout/advantages": 9.313225746154785e-09,
      "rollout/kl": 0,
      "perf/sleep_time": 11.121584177017212,
      "perf/train_time": 499.6567711830139,
      "perf/actor_train_tok_per_s": 13604.681879709395,
      "perf/update_weights_time": 1.7412362098693848,
      "perf/wait_time_ratio": 0.3129215750725044,
      "perf/step_time": 727.2194163799286,
      "perf/train_wait_time": 227.56264519691467,
      "perf/wake_up_time": 5.919387578964233,
      "perf/data_preprocess_time": 0.10505318641662598,
      "perf/actor_train_time": 393.4663851261139,
      "perf/log_probs_tflops": 129.0515051123069,
      "perf/log_probs_time": 105.58883237838745,
      "perf/actor_train_tflops": 103.89500798484565
    },
    {
      "step": 61,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 231.3719512195122,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.185185185185185,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8518518518518519,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 228.55654373880571,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7407407407407407,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.037037037037037035,
      "rollout/shopsim/success_rate": 0.87109375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6296296296296297,
      "perf/longest_effective_sample_tokens_per_sec": 24.671698594126976,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 68,
      "rollout/shopsim/strict_reward_mean": 0.871875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.871875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.06707317073170732,
      "rollout/response_len/median": 195,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.37037037037037035,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 61,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 228.55654373880571,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.2962962962962963,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 24.671698594126976,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8518518518518519,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4074074074074074,
      "perf/rollout_time": 166.02018642425537,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 231.3719512195122,
      "rollout/rollout_log_probs": -0.1446288675069809,
      "rollout/log_probs": -0.14525385200977325,
      "rollout/rewards": -0.061440500090034995,
      "rollout/total_lengths": 6310.538109756098,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.7210365853658537,
      "rollout/truncated": 0.06707317073170732,
      "rollout/returns": 1.30385160446167e-08,
      "rollout/advantages": 1.30385160446167e-08,
      "perf/actor_train_tflops": 96.81253931004228,
      "perf/data_preprocess_time": 0.09391665458679199,
      "perf/wake_up_time": 4.156991720199585,
      "perf/wait_time_ratio": 0.3082962413672196,
      "perf/log_probs_tflops": 116.9030176838363,
      "perf/step_time": 560.9932627677917,
      "perf/actor_train_tok_per_s": 13638.723477895403,
      "perf/actor_train_time": 303.5264265537262,
      "perf/sleep_time": 7.04412841796875,
      "perf/train_wait_time": 172.9521143436432,
      "perf/update_weights_time": 1.7428538799285889,
      "perf/train_time": 388.04114842414856,
      "perf/log_probs_time": 83.78786873817444
    },
    {
      "step": 62,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 229.99203187250995,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 6,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.565217391304348,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9130434782608695,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 207.49755881731141,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7391304347826086,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.28125,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.30434782608695654,
      "rollout/shopsim/success_rate": 0.8359375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6086956521739131,
      "perf/longest_effective_sample_tokens_per_sec": 19.630219902894208,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 53,
      "rollout/shopsim/strict_reward_mean": 0.8563802083333333,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8563802083333333,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.07835325365205843,
      "rollout/response_len/median": 196,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.2608695652173913,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 62,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 207.49755881731141,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.30434782608695654,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 19.630219902894208,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9565217391304348,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4782608695652174,
      "perf/rollout_time": 208.65787649154663,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 229.99203187250995,
      "rollout/rollout_log_probs": -0.15163017809391022,
      "rollout/log_probs": -0.15245940536260605,
      "rollout/rewards": -0.024042694019289764,
      "rollout/total_lengths": 7165.770252324037,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.5957060646303675,
      "rollout/truncated": 0.07835325365205843,
      "rollout/returns": -2.5902409106492996e-09,
      "rollout/advantages": -2.5902409106492996e-09,
      "perf/actor_train_tflops": 100.61553788756656,
      "perf/data_preprocess_time": 0.11236429214477539,
      "perf/wake_up_time": 6.9794628620147705,
      "perf/wait_time_ratio": 0.3131363189142431,
      "perf/log_probs_tflops": 123.82937311892562,
      "perf/step_time": 738.2783918380737,
      "perf/actor_train_tok_per_s": 13539.042532104675,
      "perf/actor_train_time": 398.53815269470215,
      "perf/sleep_time": 11.802997589111328,
      "perf/train_wait_time": 231.18177795410156,
      "perf/update_weights_time": 1.7556660175323486,
      "perf/train_time": 507.09661388397217,
      "perf/log_probs_time": 107.94189238548279
    },
    {
      "step": 63,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 262.2459259259259,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.785714285714286,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9642857142857143,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 216.6430103716929,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6071428571428571,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.39285714285714285,
      "rollout/shopsim/success_rate": 0.87890625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.8571428571428571,
      "perf/longest_effective_sample_tokens_per_sec": 20.05174154839007,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 67,
      "rollout/shopsim/strict_reward_mean": 0.8984375,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8984375,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.09481481481481481,
      "rollout/response_len/median": 221,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.35714285714285715,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 63,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 216.6430103716929,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.42857142857142855,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 20.05174154839007,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9285714285714286,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.25,
      "perf/rollout_time": 204.27153372764587,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 262.2459259259259,
      "rollout/rollout_log_probs": -0.14527519047260284,
      "rollout/log_probs": -0.146044559776783,
      "rollout/rewards": -0.1496850249502394,
      "rollout/total_lengths": 6730.939259259259,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.7185185185185186,
      "rollout/truncated": 0.09481481481481481,
      "rollout/returns": 5.587935447692871e-09,
      "rollout/advantages": 5.587935447692871e-09,
      "perf/actor_train_tflops": 98.01675361286333,
      "perf/data_preprocess_time": 0.09572172164916992,
      "perf/wake_up_time": 6.700033664703369,
      "perf/wait_time_ratio": 0.34883904636295837,
      "perf/log_probs_tflops": 122.0896024063468,
      "perf/step_time": 648.2642877101898,
      "perf/actor_train_tok_per_s": 13664.716362586243,
      "perf/actor_train_time": 332.4901797771454,
      "perf/sleep_time": 11.570079326629639,
      "perf/train_wait_time": 226.1398959159851,
      "perf/update_weights_time": 1.7234666347503662,
      "perf/train_time": 422.1243917942047,
      "perf/log_probs_time": 88.97729587554932
    },
    {
      "step": 64,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 240.5119496855346,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.375,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9166666666666666,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 201.93562485126333,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.75,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.25,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.125,
      "rollout/shopsim/success_rate": 0.8359375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.8333333333333334,
      "perf/longest_effective_sample_tokens_per_sec": 5.7959316823324105,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 46,
      "rollout/shopsim/strict_reward_mean": 0.8587239583333334,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8587239583333334,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.2540880503144654,
      "rollout/response_len/median": 197,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.2916666666666667,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 64,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/response_len/max": 1372,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 201.93562485126333,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.20833333333333334,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 5.7959316823324105,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.7916666666666666,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4583333333333333,
      "perf/rollout_time": 236.71776604652405,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 240.5119496855346,
      "rollout/rollout_log_probs": -0.1567392200231552,
      "rollout/log_probs": -0.1575736552476883,
      "rollout/rewards": -0.06426546817305703,
      "rollout/total_lengths": 8199.124528301887,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6081761006289308,
      "rollout/truncated": 0.2540880503144654,
      "rollout/returns": -1.30385160446167e-08,
      "rollout/advantages": -1.30385160446167e-08,
      "perf/actor_train_tflops": 104.77736059505423,
      "perf/data_preprocess_time": 0.15954256057739258,
      "perf/wake_up_time": 7.237299680709839,
      "perf/wait_time_ratio": 0.2958224779563865,
      "perf/log_probs_tflops": 131.22446874444816,
      "perf/step_time": 877.5659673213959,
      "perf/actor_train_tok_per_s": 13401.29204124086,
      "perf/actor_train_time": 486.3936984539032,
      "perf/sleep_time": 11.800293207168579,
      "perf/train_wait_time": 259.6037390232086,
      "perf/update_weights_time": 1.7289583683013916,
      "perf/train_time": 617.9622282981873,
      "perf/log_probs_time": 129.45514512062073
    },
    {
      "step": 65,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 234.92125984251967,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.230769230769231,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8461538461538461,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 231.19499945148186,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6153846153846154,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.07692307692307693,
      "rollout/shopsim/success_rate": 0.83984375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.7307692307692307,
      "perf/longest_effective_sample_tokens_per_sec": 8.67281526348579,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 56,
      "rollout/shopsim/strict_reward_mean": 0.83984375,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.83984375,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.03622047244094488,
      "rollout/response_len/median": 200,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.34615384615384615,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 65,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "rollout/response_len/max": 1399,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 231.19499945148186,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.2692307692307692,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 8.67281526348579,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8846153846153846,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.46153846153846156,
      "perf/rollout_time": 161.3086359500885,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 234.92125984251967,
      "rollout/rollout_log_probs": -0.14198852330446243,
      "rollout/log_probs": -0.14279961585998535,
      "rollout/rewards": 0.01388338475715457,
      "rollout/total_lengths": 6096.050393700787,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6866141732283465,
      "rollout/truncated": 0.03622047244094488,
      "rollout/returns": 9.313225746154785e-10,
      "rollout/advantages": 9.313225746154785e-10,
      "perf/actor_train_tflops": 96.87234328060048,
      "perf/data_preprocess_time": 0.09298968315124512,
      "perf/wake_up_time": 7.042602062225342,
      "perf/wait_time_ratio": 0.33953827401651504,
      "perf/log_probs_tflops": 120.49667595997028,
      "perf/step_time": 542.2307662963867,
      "perf/actor_train_tok_per_s": 13724.986808112337,
      "perf/actor_train_time": 282.03976106643677,
      "perf/sleep_time": 11.578455209732056,
      "perf/train_wait_time": 184.1080985069275,
      "perf/update_weights_time": 1.7209126949310303,
      "perf/train_time": 358.12266778945923,
      "perf/log_probs_time": 75.58120679855347
    },
    {
      "step": 66,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 250.09043250327653,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 9,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8461538461538461,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 202.64593399036667,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6923076923076923,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.15384615384615385,
      "rollout/shopsim/success_rate": 0.859375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5769230769230769,
      "perf/longest_effective_sample_tokens_per_sec": 17.399477947679042,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 10,
      "rollout/shopsim/strict_reward_mean": 0.8621651785714286,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8621651785714286,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.2712975098296199,
      "rollout/response_len/median": 209,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 66,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 202.64593399036667,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.11538461538461539,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 17.399477947679042,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8076923076923077,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4230769230769231,
      "perf/rollout_time": 235.40936183929443,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 250.09043250327653,
      "rollout/rollout_log_probs": -0.13554645329713821,
      "rollout/log_probs": -0.13628233969211578,
      "rollout/rewards": -0.043093540331014334,
      "rollout/total_lengths": 8084.6644823066845,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.5433439430818199,
      "rollout/truncated": 0.2712975098296199,
      "rollout/returns": -7.2177499532699585e-09,
      "rollout/advantages": -7.2177499532699585e-09,
      "perf/actor_train_tflops": 105.57966739787754,
      "perf/data_preprocess_time": 0.11897587776184082,
      "perf/wake_up_time": 6.953683853149414,
      "perf/wait_time_ratio": 0.30649026433745086,
      "perf/log_probs_tflops": 130.15904614434905,
      "perf/step_time": 840.5586149692535,
      "perf/actor_train_tok_per_s": 13461.092744258702,
      "perf/actor_train_time": 458.25395584106445,
      "perf/sleep_time": 11.53822112083435,
      "perf/train_wait_time": 257.6230320930481,
      "perf/update_weights_time": 1.7299914360046387,
      "perf/train_time": 582.9355828762054,
      "perf/log_probs_time": 123.90559005737305
    },
    {
      "step": 67,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 232.92985074626867,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.230769230769231,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9230769230769231,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 226.2971554987254,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6923076923076923,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.5,
      "rollout/shopsim/success_rate": 0.8984375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5769230769230769,
      "perf/longest_effective_sample_tokens_per_sec": 7.395189718533539,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 63,
      "rollout/shopsim/strict_reward_mean": 0.9033203125,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.9033203125,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.11641791044776119,
      "rollout/response_len/median": 204,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 67,
      "rollout/shopsim/all_wrong_group_rate": 0,
      "rollout/response_len/max": 1275,
      "rollout/shopsim/pending_analysis_rate": 0,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 226.2971554987254,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.07692307692307693,
      "rollout/shopsim/assisted_empty_group_rate": 0.03125,
      "perf/longest_sample_tokens_per_sec": 7.395189718533539,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.7692307692307693,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.3076923076923077,
      "perf/rollout_time": 172.40936994552612,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 232.92985074626867,
      "rollout/rollout_log_probs": -0.14002706110477448,
      "rollout/log_probs": -0.14077375084161758,
      "rollout/rewards": -0.05279379255736052,
      "rollout/total_lengths": 6657.50447761194,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.7332089552238806,
      "rollout/truncated": 0.11641791044776119,
      "rollout/returns": -3.725290298461914e-09,
      "rollout/advantages": -3.725290298461914e-09,
      "perf/actor_train_tflops": 98.69305028791699,
      "perf/data_preprocess_time": 0.09267854690551758,
      "perf/wake_up_time": 7.180835247039795,
      "perf/wait_time_ratio": 0.31919242009649407,
      "perf/log_probs_tflops": 120.63621171197299,
      "perf/step_time": 613.5821635723114,
      "perf/actor_train_tok_per_s": 13615.161234879357,
      "perf/actor_train_time": 327.61477613449097,
      "perf/sleep_time": 11.526191473007202,
      "perf/train_wait_time": 195.85077571868896,
      "perf/update_weights_time": 1.7127022743225098,
      "perf/train_time": 417.73138785362244,
      "perf/log_probs_time": 89.34106135368347
    },
    {
      "step": 68,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 235.58216560509555,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 4,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.56,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.92,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 211.71809586664352,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.64,
      "rollout/zero_std/count_0.0": 4,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.4,
      "rollout/shopsim/success_rate": 0.81640625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.72,
      "perf/longest_effective_sample_tokens_per_sec": 18.757106842942743,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 59,
      "rollout/shopsim/strict_reward_mean": 0.822265625,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.822265625,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.31337579617834393,
      "rollout/response_len/median": 197,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.32,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 68,
      "rollout/shopsim/all_wrong_group_rate": 0.125,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 211.71809586664352,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.36,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 18.757106842942743,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.84,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.36,
      "perf/rollout_time": 218.37056398391724,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 235.58216560509555,
      "rollout/rollout_log_probs": -0.13374637812376022,
      "rollout/log_probs": -0.13498559594154358,
      "rollout/rewards": -0.12258763400612363,
      "rollout/total_lengths": 8026.347770700637,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.4949044585987261,
      "rollout/truncated": 0.31337579617834393,
      "rollout/returns": 8.847564458847046e-09,
      "rollout/advantages": 8.847564458847046e-09,
      "perf/actor_train_tflops": 105.06678902760859,
      "perf/data_preprocess_time": 0.11789369583129883,
      "perf/wake_up_time": 6.919743299484253,
      "perf/wait_time_ratio": 0.28878511400857004,
      "perf/log_probs_tflops": 128.56385465852006,
      "perf/step_time": 839.1311337947845,
      "perf/actor_train_tok_per_s": 13447.704297122043,
      "perf/actor_train_time": 468.53223872184753,
      "perf/sleep_time": 11.523115634918213,
      "perf/train_wait_time": 242.3285801410675,
      "perf/update_weights_time": 1.7631785869598389,
      "perf/train_time": 596.802553653717,
      "perf/log_probs_time": 127.63353538513184
    },
    {
      "step": 69,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 265.55820476858344,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 7,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.56,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.92,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 249.7777992041009,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.8,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.32,
      "rollout/shopsim/success_rate": 0.87890625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.76,
      "perf/longest_effective_sample_tokens_per_sec": 21.613471119397015,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 9,
      "rollout/shopsim/strict_reward_mean": 0.8914620535714286,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8914620535714286,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.12201963534361851,
      "rollout/response_len/median": 202,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 69,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 249.7777992041009,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.16,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 21.613471119397015,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.84,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.36,
      "perf/rollout_time": 189.51143836975098,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 265.55820476858344,
      "rollout/rollout_log_probs": -0.14227347075939178,
      "rollout/log_probs": -0.14303823560476303,
      "rollout/rewards": -0.10099048216580342,
      "rollout/total_lengths": 6959.618513323983,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.708976157082749,
      "rollout/truncated": 0.12201963534361851,
      "rollout/returns": -9.89530235528946e-09,
      "rollout/advantages": -9.89530235528946e-09,
      "perf/actor_train_tflops": 100.19155531712804,
      "perf/data_preprocess_time": 0.10256314277648926,
      "perf/wake_up_time": 6.731716871261597,
      "perf/wait_time_ratio": 0.31251779530208484,
      "perf/log_probs_tflops": 126.47760194347966,
      "perf/step_time": 676.9642968177795,
      "perf/actor_train_tok_per_s": 13499.9998619917,
      "perf/actor_train_time": 367.57096672058105,
      "perf/sleep_time": 11.511289834976196,
      "perf/train_wait_time": 211.56338953971863,
      "perf/update_weights_time": 1.8968636989593506,
      "perf/train_time": 465.4009072780609,
      "perf/log_probs_time": 97.05936408042908
    },
    {
      "step": 70,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 259.00705052878965,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 7,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.222222222222222,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8888888888888888,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 205.10777394518726,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7407407407407407,
      "rollout/zero_std/count_0.0": 5,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.14814814814814814,
      "rollout/shopsim/success_rate": 0.79296875,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.5925925925925926,
      "perf/longest_effective_sample_tokens_per_sec": 15.246175479517944,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 73,
      "rollout/shopsim/strict_reward_mean": 0.79296875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.79296875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.15625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.22209165687426558,
      "rollout/response_len/median": 222,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.25925925925925924,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 70,
      "rollout/shopsim/all_wrong_group_rate": 0.15625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.15625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 205.10777394518726,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.25925925925925924,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 15.246175479517944,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9259259259259259,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4074074074074074,
      "perf/rollout_time": 268.6575400829315,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 259.00705052878965,
      "rollout/rollout_log_probs": -0.13560492545366287,
      "rollout/log_probs": -0.13634702563285828,
      "rollout/rewards": -0.04673719224022082,
      "rollout/total_lengths": 8333.923619271445,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.4782608695652174,
      "rollout/truncated": 0.22209165687426558,
      "rollout/returns": 2.561137080192566e-09,
      "rollout/advantages": 2.561137080192566e-09,
      "perf/actor_train_tflops": 104.37169182948226,
      "perf/data_preprocess_time": 0.12667512893676758,
      "perf/wake_up_time": 6.948782682418823,
      "perf/wait_time_ratio": 0.30345862656326555,
      "perf/log_probs_tflops": 131.69676391288516,
      "perf/step_time": 958.54598736763,
      "perf/actor_train_tok_per_s": 13439.995838795787,
      "perf/actor_train_time": 527.6913092136383,
      "perf/sleep_time": 11.369572877883911,
      "perf/train_wait_time": 290.8790488243103,
      "perf/update_weights_time": 1.7580742835998535,
      "perf/train_time": 667.6669385433197,
      "perf/log_probs_time": 139.40113401412964
    },
    {
      "step": 71,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 233.30252100840337,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.4,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 244.17285213770526,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.8,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.2,
      "rollout/shopsim/success_rate": 0.9296875,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.56,
      "perf/longest_effective_sample_tokens_per_sec": 12.270646662915622,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 77,
      "rollout/shopsim/strict_reward_mean": 0.9296875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.9296875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.10252100840336134,
      "rollout/response_len/median": 198,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.44,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 71,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/response_len/max": 1744,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 244.17285213770526,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.4,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 12.270646662915622,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.76,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.44,
      "perf/rollout_time": 142.12779879570007,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 233.30252100840337,
      "rollout/rollout_log_probs": -0.12887629866600037,
      "rollout/log_probs": -0.1299484521150589,
      "rollout/rewards": -0.13162839888524608,
      "rollout/total_lengths": 6081.759663865546,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.7815126050420168,
      "rollout/truncated": 0.10252100840336134,
      "rollout/returns": -6.05359673500061e-09,
      "rollout/advantages": -6.05359673500061e-09,
      "perf/actor_train_tflops": 98.40167140988159,
      "perf/data_preprocess_time": 0.08039069175720215,
      "perf/wake_up_time": 6.791826963424683,
      "perf/wait_time_ratio": 0.3275724486038879,
      "perf/log_probs_tflops": 120.08639880125988,
      "perf/step_time": 499.73187828063965,
      "perf/actor_train_tok_per_s": 13738.910532457008,
      "perf/actor_train_time": 263.386750459671,
      "perf/sleep_time": 11.261242151260376,
      "perf/train_wait_time": 163.6983950138092,
      "perf/update_weights_time": 1.7460737228393555,
      "perf/train_time": 336.03348326683044,
      "perf/log_probs_time": 71.94180393218994
    },
    {
      "step": 72,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 289.44101123595505,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.285714285714286,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8928571428571429,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 257.08269550363656,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6428571428571429,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.2857142857142857,
      "rollout/shopsim/success_rate": 0.77734375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.8214285714285714,
      "perf/longest_effective_sample_tokens_per_sec": 13.62578288620252,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 60,
      "rollout/shopsim/strict_reward_mean": 0.78759765625,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.78759765625,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.33052434456928836,
      "rollout/response_len/median": 220.5,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.21428571428571427,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 72,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 257.08269550363656,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.17857142857142858,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 13.62578288620252,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8928571428571429,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.35714285714285715,
      "perf/rollout_time": 300.60658049583435,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 289.44101123595505,
      "rollout/rollout_log_probs": -0.1352050006389618,
      "rollout/log_probs": -0.13608553260564804,
      "rollout/rewards": -0.12013906246044216,
      "rollout/total_lengths": 9831.201310861423,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.40554775280898875,
      "rollout/truncated": 0.33052434456928836,
      "rollout/returns": -4.6566128730773926e-09,
      "rollout/advantages": -4.6566128730773926e-09,
      "perf/actor_train_tflops": 107.24722244421683,
      "perf/data_preprocess_time": 0.17517948150634766,
      "perf/wake_up_time": 6.6545562744140625,
      "perf/wait_time_ratio": 0.24451592835131625,
      "perf/log_probs_tflops": 134.56204848761183,
      "perf/step_time": 1322.1463577747345,
      "perf/actor_train_tok_per_s": 13314.823681485666,
      "perf/actor_train_time": 788.5739421844482,
      "perf/sleep_time": 11.51755166053772,
      "perf/train_wait_time": 323.2858440876007,
      "perf/update_weights_time": 1.735121250152588,
      "perf/train_time": 998.8605136871338,
      "perf/log_probs_time": 209.50029110908508
    },
    {
      "step": 73,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 273.1438450899032,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.62962962962963,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8518518518518519,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 196.65363510151997,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6666666666666666,
      "rollout/zero_std/count_0.0": 3,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.15625,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.4444444444444444,
      "rollout/shopsim/success_rate": 0.89453125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6296296296296297,
      "perf/longest_effective_sample_tokens_per_sec": 16.31519248493948,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 54,
      "rollout/shopsim/strict_reward_mean": 0.89453125,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.89453125,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.09375,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.18118948824343015,
      "rollout/response_len/median": 202,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.37037037037037035,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 73,
      "rollout/shopsim/all_wrong_group_rate": 0.09375,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.09375,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 196.65363510151997,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.4074074074074074,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 16.31519248493948,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8888888888888888,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.37037037037037035,
      "perf/rollout_time": 251.05434727668762,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 273.1438450899032,
      "rollout/rollout_log_probs": -0.1268034242093563,
      "rollout/log_probs": -0.128240704536438,
      "rollout/rewards": -0.08362027190697803,
      "rollout/total_lengths": 7269.174273858921,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.607192254495159,
      "rollout/truncated": 0.18118948824343015,
      "rollout/returns": 1.862645149230957e-09,
      "rollout/advantages": 1.862645149230957e-09,
      "perf/actor_train_tflops": 100.84281475969097,
      "perf/data_preprocess_time": 0.10316085815429688,
      "perf/wake_up_time": 7.666072130203247,
      "perf/wait_time_ratio": 0.35661566611851914,
      "perf/log_probs_tflops": 121.78750039574392,
      "perf/step_time": 767.4950966835022,
      "perf/actor_train_tok_per_s": 13599.396207726706,
      "perf/actor_train_time": 386.45928978919983,
      "perf/sleep_time": 11.201563835144043,
      "perf/train_wait_time": 273.7007751464844,
      "perf/update_weights_time": 1.7479698657989502,
      "perf/train_time": 493.7943215370178,
      "perf/log_probs_time": 106.66568803787231
    },
    {
      "step": 74,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 221.80492091388402,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 13,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.689655172413793,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8620689655172413,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 197.67387049956977,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7241379310344828,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.09375,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.41379310344827586,
      "rollout/shopsim/success_rate": 0.96484375,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6896551724137931,
      "perf/longest_effective_sample_tokens_per_sec": 5.425549196245135,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 45,
      "rollout/shopsim/strict_reward_mean": 0.96484375,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.96484375,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.01757469244288225,
      "rollout/response_len/median": 179,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.2413793103448276,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 74,
      "rollout/shopsim/all_wrong_group_rate": 0,
      "rollout/response_len/max": 866,
      "rollout/shopsim/pending_analysis_rate": 0,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 197.67387049956977,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.3793103448275862,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 5.425549196245135,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.9310344827586207,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.4482758620689655,
      "perf/rollout_time": 159.61517786979675,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 221.80492091388402,
      "rollout/rollout_log_probs": -0.11989164352416992,
      "rollout/log_probs": -0.12065617367625237,
      "rollout/rewards": -0.030446514042064887,
      "rollout/total_lengths": 5537.753954305799,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.8892794376098418,
      "rollout/truncated": 0.01757469244288225,
      "rollout/returns": 9.313225746154785e-10,
      "rollout/advantages": 9.313225746154785e-10,
      "perf/actor_train_tflops": 94.23551852483186,
      "perf/data_preprocess_time": 0.07480883598327637,
      "perf/wake_up_time": 6.819113254547119,
      "perf/wait_time_ratio": 0.38544176439530314,
      "perf/log_probs_tflops": 114.17308061780459,
      "perf/step_time": 469.6946506500244,
      "perf/actor_train_tok_per_s": 13956.890246986106,
      "perf/actor_train_time": 225.76533484458923,
      "perf/sleep_time": 11.052287101745605,
      "perf/train_wait_time": 181.03993487358093,
      "perf/update_weights_time": 1.6970603466033936,
      "perf/train_time": 288.6547157764435,
      "perf/log_probs_time": 62.11362981796265
    },
    {
      "step": 75,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 246.7919826652221,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 8,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 3.88,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.76,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 261.91844921187146,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.28,
      "rollout/shopsim/success_rate": 0.90625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.76,
      "perf/longest_effective_sample_tokens_per_sec": 18.83880201364992,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 58,
      "rollout/shopsim/strict_reward_mean": 0.91015625,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.91015625,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.1722643553629469,
      "rollout/response_len/median": 197,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.08,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 75,
      "rollout/shopsim/all_wrong_group_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 261.91844921187146,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.32,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 18.83880201364992,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.84,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.24,
      "perf/rollout_time": 217.4235918521881,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 246.7919826652221,
      "rollout/rollout_log_probs": -0.12454785779118538,
      "rollout/log_probs": -0.1261356621980667,
      "rollout/rewards": -0.19269506076304393,
      "rollout/total_lengths": 8387.589382448537,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6175514626218852,
      "rollout/truncated": 0.1722643553629469,
      "rollout/returns": 8.847564458847046e-09,
      "rollout/advantages": 8.847564458847046e-09,
      "perf/actor_train_tflops": 103.37624385624525,
      "perf/data_preprocess_time": 0.13314127922058105,
      "perf/wake_up_time": 6.807807683944702,
      "perf/wait_time_ratio": 0.24567575988446425,
      "perf/log_probs_tflops": 128.60675768706736,
      "perf/step_time": 974.9976196289062,
      "perf/actor_train_tok_per_s": 13356.74163170278,
      "perf/actor_train_time": 579.6132929325104,
      "perf/sleep_time": 11.382997274398804,
      "perf/train_wait_time": 239.53328108787537,
      "perf/update_weights_time": 1.7171297073364258,
      "perf/train_time": 735.4643385410309,
      "perf/log_probs_time": 155.30092453956604
    },
    {
      "step": 76,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 272.7401847575058,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 5,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.5,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9642857142857143,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 246.96466479813532,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6785714285714286,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.125,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.35714285714285715,
      "rollout/shopsim/success_rate": 0.85546875,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.7142857142857143,
      "perf/longest_effective_sample_tokens_per_sec": 17.131198079759557,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 62,
      "rollout/shopsim/strict_reward_mean": 0.8671875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.8671875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.20092378752886836,
      "rollout/response_len/median": 198.5,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.32142857142857145,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 76,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 246.96466479813532,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.2857142857142857,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 17.131198079759557,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8571428571428571,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.32142857142857145,
      "perf/rollout_time": 239.09594535827637,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 272.7401847575058,
      "rollout/rollout_log_probs": -0.12606461718678474,
      "rollout/log_probs": -0.12748649716377258,
      "rollout/rewards": -0.06152621637728693,
      "rollout/total_lengths": 7856.645496535797,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6374133949191686,
      "rollout/truncated": 0.20092378752886836,
      "rollout/returns": 0,
      "rollout/advantages": 0,
      "perf/actor_train_tflops": 101.59211926164863,
      "perf/data_preprocess_time": 0.12386870384216309,
      "perf/wake_up_time": 7.520815134048462,
      "perf/wait_time_ratio": 0.28840508852388147,
      "perf/log_probs_tflops": 124.03824605850811,
      "perf/step_time": 912.2639644145966,
      "perf/actor_train_tok_per_s": 13359.921773523838,
      "perf/actor_train_time": 509.27356576919556,
      "perf/sleep_time": 12.665958166122437,
      "perf/train_wait_time": 263.1015694141388,
      "perf/update_weights_time": 1.713705062866211,
      "perf/train_time": 649.1623950004578,
      "perf/log_probs_time": 139.0382468700409
    },
    {
      "step": 77,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 267.6924066924067,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 10,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.346153846153846,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.8461538461538461,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 213.69642950593644,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.6923076923076923,
      "rollout/zero_std/count_0.0": 2,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.1875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.38461538461538464,
      "rollout/shopsim/success_rate": 0.89453125,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.6153846153846154,
      "perf/longest_effective_sample_tokens_per_sec": 16.83294615319097,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 3,
      "rollout/shopsim/strict_reward_mean": 0.89453125,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.89453125,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.16859716859716858,
      "rollout/response_len/median": 203,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.23076923076923078,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 77,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 213.69642950593644,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.19230769230769232,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 16.83294615319097,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 1,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.38461538461538464,
      "perf/rollout_time": 243.33232951164246,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 267.6924066924067,
      "rollout/rollout_log_probs": -0.126546211540699,
      "rollout/log_probs": -0.12755507230758667,
      "rollout/rewards": -0.03130268659370746,
      "rollout/total_lengths": 7602.929214929215,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6602316602316602,
      "rollout/truncated": 0.16859716859716858,
      "rollout/returns": 3.0850060284137726e-09,
      "rollout/advantages": 3.0850060284137726e-09,
      "perf/actor_train_tflops": 100.9413399088107,
      "perf/data_preprocess_time": 0.11426949501037598,
      "perf/wake_up_time": 7.2688210010528564,
      "perf/wait_time_ratio": 0.32261504981683553,
      "perf/log_probs_tflops": 123.94745339532398,
      "perf/step_time": 830.1242394447327,
      "perf/actor_train_tok_per_s": 13374.324572435393,
      "perf/actor_train_time": 441.7027542591095,
      "perf/sleep_time": 13.893450736999512,
      "perf/train_wait_time": 267.8105728626251,
      "perf/update_weights_time": 1.7145907878875732,
      "perf/train_time": 562.3136665821075,
      "perf/log_probs_time": 119.90583276748657
    },
    {
      "step": 78,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 230.60723860589812,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 12,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.76,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.76,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 196.37651208145553,
      "rollout/zero_std/count_0.5": 1,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.56,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.4,
      "rollout/shopsim/success_rate": 0.9140625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.64,
      "perf/longest_effective_sample_tokens_per_sec": 6.013447801556143,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 6,
      "rollout/shopsim/strict_reward_mean": 0.9296875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.9296875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.0625,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.18230563002680966,
      "rollout/response_len/median": 187.5,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.4,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 78,
      "rollout/shopsim/all_wrong_group_rate": 0.0625,
      "rollout/response_len/max": 1317,
      "rollout/shopsim/pending_analysis_rate": 0.0625,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 196.37651208145553,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.52,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 6.013447801556143,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.96,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.52,
      "perf/rollout_time": 219.0091347694397,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 230.60723860589812,
      "rollout/rollout_log_probs": -0.1240248791873455,
      "rollout/log_probs": -0.12484591454267502,
      "rollout/rewards": -0.07511155023651532,
      "rollout/total_lengths": 7602.032171581769,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.703083109919571,
      "rollout/truncated": 0.18230563002680966,
      "rollout/returns": -4.6566128730773926e-09,
      "rollout/advantages": -4.6566128730773926e-09,
      "perf/actor_train_tflops": 101.18412996249214,
      "perf/data_preprocess_time": 0.10632181167602539,
      "perf/wake_up_time": 6.805798768997192,
      "perf/wait_time_ratio": 0.3076357286496534,
      "perf/log_probs_tflops": 122.07603546237269,
      "perf/step_time": 785.7272098064423,
      "perf/actor_train_tok_per_s": 13324.772112345678,
      "perf/actor_train_time": 425.60697865486145,
      "perf/sleep_time": 12.376241445541382,
      "perf/train_wait_time": 241.71776270866394,
      "perf/update_weights_time": 1.7158095836639404,
      "perf/train_time": 544.0094470977783,
      "perf/log_probs_time": 117.58975100517273
    },
    {
      "step": 79,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 210.86363636363637,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 11,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.44,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.92,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 219.80702532234835,
      "rollout/shopsim/full_skill_retry_success_rate": 0,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.44,
      "rollout/zero_std/count_0.0": 1,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.21875,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.24,
      "rollout/shopsim/success_rate": 0.90625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.8,
      "perf/longest_effective_sample_tokens_per_sec": 27.725481961024197,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 56,
      "rollout/shopsim/strict_reward_mean": 0.9085286458333334,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.9085286458333334,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0.03125,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.00487012987012987,
      "rollout/response_len/median": 175.5,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.44,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 79,
      "rollout/shopsim/all_wrong_group_rate": 0.03125,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0.03125,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 219.80702532234835,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.44,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 27.725481961024197,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.88,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.28,
      "perf/rollout_time": 147.73413157463074,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 210.86363636363637,
      "rollout/rollout_log_probs": -0.10677902773022652,
      "rollout/log_probs": -0.10759465023875237,
      "rollout/rewards": -0.02945723065308162,
      "rollout/total_lengths": 5500.887987012987,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.7468885281385281,
      "rollout/truncated": 0.00487012987012987,
      "rollout/returns": 5.587935447692871e-09,
      "rollout/advantages": 5.587935447692871e-09,
      "perf/actor_train_tflops": 93.45292901607006,
      "perf/data_preprocess_time": 0.08683466911315918,
      "perf/wake_up_time": 7.385131597518921,
      "perf/wait_time_ratio": 0.3528892392519517,
      "perf/log_probs_tflops": 114.61959187344846,
      "perf/step_time": 484.91418719291687,
      "perf/actor_train_tok_per_s": 13768.323418306927,
      "perf/actor_train_time": 246.11181020736694,
      "perf/sleep_time": 12.336070775985718,
      "perf/train_wait_time": 171.12099862098694,
      "perf/update_weights_time": 1.7113370895385742,
      "perf/train_time": 313.79318857192993,
      "perf/log_probs_time": 66.88754558563232
    },
    {
      "step": 80,
      "round": 3,
      "run_id": "y4xk6byf",
      "rollout/shopsim/chunk/online-214dfd1028dc4157/target_rho": 0.3190629307171016,
      "rollout/response_len/mean": 247.21917808219177,
      "rollout/shopsim/internalization_deficit_rate": 0,
      "rollout/zero_std/count_1.0": 9,
      "rollout/shopsim/zero_std_group_drop_rate": 0.5,
      "rollout/dynamic_filter/drop_shopsim_zero_std_group": 16,
      "rollout/shopsim/zero_std_group_drop_count": 16,
      "rollout/repetition_frac": 0,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/target_rho": 0.7037147129653498,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/target_rho": 0.29279886960426194,
      "rollout/shopsim/rho_max": 0.8999999999999999,
      "rollout/shopsim/skills_per_assisted_group_mean": 4.428571428571429,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/inclusion_rate": 0.9047619047619048,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/target_rho": 0.8999999999999999,
      "rollout/shopsim/trajectory_count": 256,
      "perf/tokens_per_gpu_per_sec": 255.9275698393503,
      "rollout/shopsim/q": 0.2,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/inclusion_rate": 0.7619047619047619,
      "rollout/shopsim/scored_rate": 1,
      "rollout/shopsim/skill_free_group_rate": 0.34375,
      "rollout/shopsim/chunk/online-3297dc4c0f4f6109/inclusion_rate": 0.38095238095238093,
      "rollout/shopsim/success_rate": 0.9140625,
      "rollout/shopsim/chunk/online-8e3e179b9270814a/inclusion_rate": 0.8095238095238095,
      "perf/longest_effective_sample_tokens_per_sec": 21.122207691270347,
      "rollout/shopsim/train_trajectory_count": 128,
      "rollout/response_len/min": 49,
      "rollout/shopsim/strict_reward_mean": 0.921875,
      "rollout/shopsim/rho_mean": 0.5663062215285343,
      "rollout/shopsim/reward_mean": 0.921875,
      "rollout/prefix_cache_hit_rate": 0,
      "rollout/shopsim/full_skill_retry_rate": 0,
      "rollout/shopsim/train_group_count": 16,
      "rollout/truncated_ratio": 0.22291407222914073,
      "rollout/response_len/median": 187,
      "rollout/shopsim/chunk/online-214dfd1028dc4157/inclusion_rate": 0.2857142857142857,
      "rollout/shopsim/chunk/chunk-a9e98e128a35/target_rho": 0.8876413579359135,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/target_rho": 0.3993089719266536,
      "rollout/step": 80,
      "rollout/shopsim/all_wrong_group_rate": 0,
      "rollout/response_len/max": 4096,
      "rollout/shopsim/pending_analysis_rate": 0,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/target_rho": 0.3415999358559877,
      "rollout/shopsim/group_count": 32,
      "rollout/shopsim/chunk/chunk-1f8d27322e9f/target_rho": 0.6863229932230063,
      "perf/effective_tokens_per_gpu_per_sec": 255.9275698393503,
      "rollout/shopsim/chunk/online-3d2cdd6a2aac9717/inclusion_rate": 0.14285714285714285,
      "rollout/shopsim/assisted_empty_group_rate": 0,
      "perf/longest_sample_tokens_per_sec": 21.122207691270347,
      "rollout/shopsim/chunk/chunk-c83ed5ae8b70/inclusion_rate": 0.8571428571428571,
      "rollout/shopsim/chunk/online-ea5dd903b14ec0fc/inclusion_rate": 0.2857142857142857,
      "perf/rollout_time": 193.9191234111786,
      "rollout/avg_cached_tokens_per_sample": 0,
      "rollout/shopsim/rho_min": 0.29279886960426194,
      "rollout/shopsim/runtime_error_group_rate": 0,
      "rollout/response_lengths": 247.21917808219177,
      "rollout/rollout_log_probs": -0.11619194969534874,
      "rollout/log_probs": -0.11700985208153725,
      "rollout/rewards": -0.10488933295122861,
      "rollout/total_lengths": 7696.4819427148195,
      "rollout/kl": 0,
      "rollout/raw_reward": 0.6525529265255293,
      "rollout/truncated": 0.22291407222914073,
      "rollout/returns": -1.862645149230957e-09,
      "rollout/advantages": -1.862645149230957e-09,
      "perf/actor_train_tflops": 102.43451483914053,
      "perf/data_preprocess_time": 0.11639070510864258,
      "perf/wake_up_time": 7.427615165710449,
      "perf/wait_time_ratio": 0.2690706044945577,
      "perf/log_probs_tflops": 125.36374718627475,
      "perf/step_time": 808.7974395751953,
      "perf/actor_train_tok_per_s": 13315.303629806138,
      "perf/actor_train_time": 464.14825916290283,
      "perf/sleep_time": 12.321182250976562,
      "perf/train_wait_time": 217.62361598014832,
      "perf/update_weights_time": 1.7111709117889404,
      "perf/train_time": 591.173823595047,
      "perf/log_probs_time": 126.41826367378235
    }
  ],
  "updates": [
    {
      "update": 1,
      "train_step": 2,
      "step": 1,
      "round": 0,
      "train/loss": -1.5832483768463135e-08,
      "train/step": 2,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5716050934419454,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18806378543376923,
      "train/pg_loss": -1.5832483768463135e-08,
      "train/train_rollout_logprob_abs_diff": 0.009453191421926022,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 2,
      "train_step": 3,
      "step": 1,
      "round": 0,
      "train/loss": 0.0001513175666332245,
      "train/step": 3,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5798214396245585,
      "train/ppo_kl": -4.9009613576345146e-06,
      "train/entropy_loss": 0.1946052759885788,
      "train/pg_loss": 0.0001513175666332245,
      "train/train_rollout_logprob_abs_diff": 0.010037851519882679,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.001969448523595929
    },
    {
      "update": 3,
      "train_step": 4,
      "step": 2,
      "round": 0,
      "train/loss": -1.3969838619232178e-08,
      "train/step": 4,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.6049983688298668,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1950695812702179,
      "train/pg_loss": -1.3969838619232178e-08,
      "train/train_rollout_logprob_abs_diff": 0.011565715074539185,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 4,
      "train_step": 5,
      "step": 2,
      "round": 0,
      "train/loss": 0.00022393465042114258,
      "train/step": 5,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.000693916103875,
      "train/ppo_kl": 4.533681203611195e-05,
      "train/entropy_loss": 0.21762660145759583,
      "train/pg_loss": 0.00022393465042114258,
      "train/train_rollout_logprob_abs_diff": 0.011645268648862839,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.002885817550122738
    },
    {
      "update": 5,
      "train_step": 6,
      "step": 3,
      "round": 0,
      "train/loss": 7.450580596923828e-09,
      "train/step": 6,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8467069287966291,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.20660385489463806,
      "train/pg_loss": 7.450580596923828e-09,
      "train/train_rollout_logprob_abs_diff": 0.01066572591662407,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 6,
      "train_step": 7,
      "step": 3,
      "round": 0,
      "train/loss": 0.0005662236362695694,
      "train/step": 7,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7945008889379714,
      "train/ppo_kl": 0.000541273970156908,
      "train/entropy_loss": 0.2019944190979004,
      "train/pg_loss": 0.0005662236362695694,
      "train/train_rollout_logprob_abs_diff": 0.01091139018535614,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0041798921301960945
    },
    {
      "update": 7,
      "train_step": 8,
      "step": 4,
      "round": 0,
      "train/loss": -1.30385160446167e-08,
      "train/step": 8,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.680135835651098,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18698650598526,
      "train/pg_loss": -1.30385160446167e-08,
      "train/train_rollout_logprob_abs_diff": 0.010189298540353775,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 8,
      "train_step": 9,
      "step": 4,
      "round": 0,
      "train/loss": 0.0001058848574757576,
      "train/step": 9,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.245910918117503,
      "train/ppo_kl": -4.820346657652408e-05,
      "train/entropy_loss": 0.1880761682987213,
      "train/pg_loss": 0.0001058848574757576,
      "train/train_rollout_logprob_abs_diff": 0.010209959000349045,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0008586939075030386
    },
    {
      "update": 9,
      "train_step": 10,
      "step": 5,
      "round": 0,
      "train/loss": 1.4901161193847656e-08,
      "train/step": 10,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.900806910682247,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18700279295444489,
      "train/pg_loss": 1.4901161193847656e-08,
      "train/train_rollout_logprob_abs_diff": 0.010394353419542313,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 10,
      "train_step": 11,
      "step": 5,
      "round": 0,
      "train/loss": 0.0004104822874069214,
      "train/step": 11,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8072337893435626,
      "train/ppo_kl": 0.00020537880482152104,
      "train/entropy_loss": 0.17268969118595123,
      "train/pg_loss": 0.0004104822874069214,
      "train/train_rollout_logprob_abs_diff": 0.009429998695850372,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0021267938427627087
    },
    {
      "update": 11,
      "train_step": 12,
      "step": 6,
      "round": 0,
      "train/loss": -1.1175870895385742e-08,
      "train/step": 12,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7544105671320425,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.19069011509418488,
      "train/pg_loss": -1.1175870895385742e-08,
      "train/train_rollout_logprob_abs_diff": 0.010394438169896603,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 12,
      "train_step": 13,
      "step": 6,
      "round": 0,
      "train/loss": 0.0005107056349515915,
      "train/step": 13,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.877149113198622,
      "train/ppo_kl": 0.0007192570483312011,
      "train/entropy_loss": 0.17904925346374512,
      "train/pg_loss": 0.0005107056349515915,
      "train/train_rollout_logprob_abs_diff": 0.009693820029497147,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0028704700525850058
    },
    {
      "update": 13,
      "train_step": 14,
      "step": 7,
      "round": 0,
      "train/loss": 0,
      "train/step": 14,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.0887227982690844,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.2002553939819336,
      "train/pg_loss": 0,
      "train/train_rollout_logprob_abs_diff": 0.010787541046738625,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 14,
      "train_step": 15,
      "step": 7,
      "round": 0,
      "train/loss": 8.938554674386978e-05,
      "train/step": 15,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8086843117765876,
      "train/ppo_kl": 0.0002471660263836384,
      "train/entropy_loss": 0.18212303519248962,
      "train/pg_loss": 8.938554674386978e-05,
      "train/train_rollout_logprob_abs_diff": 0.010171739384531975,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.002472338732331991
    },
    {
      "update": 15,
      "train_step": 16,
      "step": 8,
      "round": 0,
      "train/loss": -1.6298145055770874e-08,
      "train/step": 16,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.564909290843941,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.19825954735279083,
      "train/pg_loss": -1.6298145055770874e-08,
      "train/train_rollout_logprob_abs_diff": 0.011085400357842445,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 16,
      "train_step": 17,
      "step": 8,
      "round": 0,
      "train/loss": -0.00010277703404426575,
      "train/step": 17,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5894886716935155,
      "train/ppo_kl": 0.0005067376187071204,
      "train/entropy_loss": 0.19235742092132568,
      "train/pg_loss": -0.00010277703404426575,
      "train/train_rollout_logprob_abs_diff": 0.010631348937749863,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.002468240913003683
    },
    {
      "update": 17,
      "train_step": 18,
      "step": 9,
      "round": 0,
      "train/loss": -5.587935447692871e-09,
      "train/step": 18,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.0928136525315417,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18229128420352936,
      "train/pg_loss": -5.587935447692871e-09,
      "train/train_rollout_logprob_abs_diff": 0.010913250967860222,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 18,
      "train_step": 19,
      "step": 9,
      "round": 0,
      "train/loss": 0.00021376041695475578,
      "train/step": 19,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5483985872893415,
      "train/ppo_kl": 0.00015007896581664681,
      "train/entropy_loss": 0.16618984937667847,
      "train/pg_loss": 0.00021376041695475578,
      "train/train_rollout_logprob_abs_diff": 0.00989993754774332,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0022167502902448177
    },
    {
      "update": 19,
      "train_step": 20,
      "step": 10,
      "round": 0,
      "train/loss": 7.450580596923828e-09,
      "train/step": 20,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7827532264424573,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17033249139785767,
      "train/pg_loss": 7.450580596923828e-09,
      "train/train_rollout_logprob_abs_diff": 0.009500138461589813,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 20,
      "train_step": 21,
      "step": 10,
      "round": 0,
      "train/loss": -0.007292289286851883,
      "train/step": 21,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.5544846013774123,
      "train/ppo_kl": 0.00017803606169763952,
      "train/entropy_loss": 0.17362499237060547,
      "train/pg_loss": -0.007292289286851883,
      "train/train_rollout_logprob_abs_diff": 0.009826273657381535,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.002416501985862851
    },
    {
      "update": 21,
      "train_step": 22,
      "step": 11,
      "round": 0,
      "train/loss": -1.862645149230957e-09,
      "train/step": 22,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.376647076929511,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18104249238967896,
      "train/pg_loss": -1.862645149230957e-09,
      "train/train_rollout_logprob_abs_diff": 0.01059776358306408,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 22,
      "train_step": 23,
      "step": 11,
      "round": 0,
      "train/loss": 0.00028558261692523956,
      "train/step": 23,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.3628057332491283,
      "train/ppo_kl": 0.0003082345938310027,
      "train/entropy_loss": 0.1873510777950287,
      "train/pg_loss": 0.00028558261692523956,
      "train/train_rollout_logprob_abs_diff": 0.010438353754580021,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0029210965149104595
    },
    {
      "update": 23,
      "train_step": 24,
      "step": 12,
      "round": 0,
      "train/loss": -1.862645149230957e-09,
      "train/step": 24,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 6.063177326551887,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.16372162103652954,
      "train/pg_loss": -1.862645149230957e-09,
      "train/train_rollout_logprob_abs_diff": 0.00999264232814312,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 24,
      "train_step": 25,
      "step": 12,
      "round": 0,
      "train/loss": 0.00016181915998458862,
      "train/step": 25,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8450303239062016,
      "train/ppo_kl": 0.00031598383793607354,
      "train/entropy_loss": 0.1784619688987732,
      "train/pg_loss": 0.00016181915998458862,
      "train/train_rollout_logprob_abs_diff": 0.010152394883334637,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.002379489364102483
    },
    {
      "update": 25,
      "train_step": 26,
      "step": 13,
      "round": 0,
      "train/loss": -7.450580596923828e-09,
      "train/step": 26,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.0305415165049556,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17299027740955353,
      "train/pg_loss": -7.450580596923828e-09,
      "train/train_rollout_logprob_abs_diff": 0.010025481693446636,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 26,
      "train_step": 27,
      "step": 13,
      "round": 0,
      "train/loss": 0.0002325214445590973,
      "train/step": 27,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.1896283000368373,
      "train/ppo_kl": 0.0002036926307482645,
      "train/entropy_loss": 0.18496987223625183,
      "train/pg_loss": 0.0002325214445590973,
      "train/train_rollout_logprob_abs_diff": 0.010585332289338112,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0012768437154591084
    },
    {
      "update": 27,
      "train_step": 28,
      "step": 14,
      "round": 0,
      "train/loss": 4.0978193283081055e-08,
      "train/step": 28,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.4332674888516155,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17240190505981445,
      "train/pg_loss": 4.0978193283081055e-08,
      "train/train_rollout_logprob_abs_diff": 0.00964379496872425,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 28,
      "train_step": 29,
      "step": 14,
      "round": 0,
      "train/loss": 0.002470601350069046,
      "train/step": 29,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 9.967743921483805,
      "train/ppo_kl": 0.00025858383742161095,
      "train/entropy_loss": 0.1668476164340973,
      "train/pg_loss": 0.002470601350069046,
      "train/train_rollout_logprob_abs_diff": 0.009788467548787594,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0023342594504356384
    },
    {
      "update": 29,
      "train_step": 30,
      "step": 15,
      "round": 0,
      "train/loss": -2.9802322387695312e-08,
      "train/step": 30,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.344242807076838,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17494012415409088,
      "train/pg_loss": -2.9802322387695312e-08,
      "train/train_rollout_logprob_abs_diff": 0.009843398816883564,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 30,
      "train_step": 31,
      "step": 15,
      "round": 0,
      "train/loss": 0.0005398709326982498,
      "train/step": 31,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.334417965249735,
      "train/ppo_kl": 0.0005076018278487027,
      "train/entropy_loss": 0.1813342273235321,
      "train/pg_loss": 0.0005398709326982498,
      "train/train_rollout_logprob_abs_diff": 0.01051272451877594,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0035054918844252825
    },
    {
      "update": 31,
      "train_step": 32,
      "step": 16,
      "round": 0,
      "train/loss": -1.1175870895385742e-08,
      "train/step": 32,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.9496555488757712,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17412276566028595,
      "train/pg_loss": -1.1175870895385742e-08,
      "train/train_rollout_logprob_abs_diff": 0.009888405911624432,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 32,
      "train_step": 33,
      "step": 16,
      "round": 0,
      "train/loss": -6.795302033424377e-05,
      "train/step": 33,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7402996560470627,
      "train/ppo_kl": 0.00014324506628327072,
      "train/entropy_loss": 0.17809121310710907,
      "train/pg_loss": -6.795302033424377e-05,
      "train/train_rollout_logprob_abs_diff": 0.010547874495387077,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0013971212320029736
    },
    {
      "update": 33,
      "train_step": 34,
      "step": 17,
      "round": 0,
      "train/loss": 9.313225746154785e-09,
      "train/step": 34,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.9654226493083946,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.17317134141921997,
      "train/pg_loss": 9.313225746154785e-09,
      "train/train_rollout_logprob_abs_diff": 0.009906629100441933,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 34,
      "train_step": 35,
      "step": 17,
      "round": 0,
      "train/loss": 0.00019548926502466202,
      "train/step": 35,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7859485962957724,
      "train/ppo_kl": 0.00030705248354934156,
      "train/entropy_loss": 0.159990131855011,
      "train/pg_loss": 0.00019548926502466202,
      "train/train_rollout_logprob_abs_diff": 0.00950855016708374,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0029927515424788
    },
    {
      "update": 35,
      "train_step": 36,
      "step": 18,
      "round": 0,
      "train/loss": 8.381903171539307e-09,
      "train/step": 36,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7434617928950757,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.18171241879463196,
      "train/pg_loss": 8.381903171539307e-09,
      "train/train_rollout_logprob_abs_diff": 0.010481154546141624,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 36,
      "train_step": 37,
      "step": 18,
      "round": 0,
      "train/loss": 0.00592871755361557,
      "train/step": 37,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7184746261556982,
      "train/ppo_kl": 0.002759676892310381,
      "train/entropy_loss": 0.17726729810237885,
      "train/pg_loss": 0.00592871755361557,
      "train/train_rollout_logprob_abs_diff": 0.010782839730381966,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0024944066535681486
    },
    {
      "update": 37,
      "train_step": 38,
      "step": 19,
      "round": 0,
      "train/loss": -1.0244548320770264e-08,
      "train/step": 38,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8428289408125855,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1600552797317505,
      "train/pg_loss": -1.0244548320770264e-08,
      "train/train_rollout_logprob_abs_diff": 0.009782875888049603,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 38,
      "train_step": 39,
      "step": 19,
      "round": 0,
      "train/loss": 0.00021580979228019714,
      "train/step": 39,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.611587632985406,
      "train/ppo_kl": 0.0031463007908314466,
      "train/entropy_loss": 0.15717901289463043,
      "train/pg_loss": 0.00021580979228019714,
      "train/train_rollout_logprob_abs_diff": 0.010231465101242065,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0018037890549749136
    },
    {
      "update": 39,
      "train_step": 40,
      "step": 20,
      "round": 0,
      "train/loss": 3.725290298461914e-09,
      "train/step": 40,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.6475869104039988,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1692807376384735,
      "train/pg_loss": 3.725290298461914e-09,
      "train/train_rollout_logprob_abs_diff": 0.01032445952296257,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0
    },
    {
      "update": 40,
      "train_step": 41,
      "step": 20,
      "round": 0,
      "train/loss": 0.0002836789935827255,
      "train/step": 41,
      "train/lr-pg_1": 1e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.67925998658764,
      "train/ppo_kl": 0.0017486463766545057,
      "train/entropy_loss": 0.14106778800487518,
      "train/pg_loss": 0.0002836789935827255,
      "train/train_rollout_logprob_abs_diff": 0.008858976885676384,
      "train/global_batch_size": 64,
      "train/pg_clipfrac": 0.0035239316057413816
    },
    {
      "update": 41,
      "train_step": 42,
      "step": 21,
      "round": 1,
      "train/pg_loss": -8.381903171539307e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.6840553447481483,
      "train/loss": -8.381903171539307e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 42,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.011935489252209663,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.16218498349189758
    },
    {
      "update": 42,
      "train_step": 43,
      "step": 21,
      "round": 1,
      "train/pg_loss": 0.00038815662264823914,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.2021021272368047,
      "train/loss": 0.00038815662264823914,
      "train/lr-pg_1": 1e-06,
      "train/step": 43,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009625840932130814,
      "train/pg_clipfrac": 0.001988448202610016,
      "train/ppo_kl": 0.002770230872556567,
      "train/entropy_loss": 0.15296417474746704
    },
    {
      "update": 43,
      "train_step": 44,
      "step": 22,
      "round": 1,
      "train/pg_loss": 1.30385160446167e-08,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.715868736977637,
      "train/loss": 1.30385160446167e-08,
      "train/lr-pg_1": 1e-06,
      "train/step": 44,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010732773691415787,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1706748604774475
    },
    {
      "update": 44,
      "train_step": 45,
      "step": 22,
      "round": 1,
      "train/pg_loss": 0.00032370537519454956,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.9525988671234356,
      "train/loss": 0.00032370537519454956,
      "train/lr-pg_1": 1e-06,
      "train/step": 45,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.00996735505759716,
      "train/pg_clipfrac": 0.0027661737985908985,
      "train/ppo_kl": 0.00018472809460945427,
      "train/entropy_loss": 0.1560744047164917
    },
    {
      "update": 45,
      "train_step": 46,
      "step": 23,
      "round": 1,
      "train/pg_loss": -5.005858838558197e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.1977533090856645,
      "train/loss": -5.005858838558197e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 46,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009817853569984436,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15600386261940002
    },
    {
      "update": 46,
      "train_step": 47,
      "step": 23,
      "round": 1,
      "train/pg_loss": -2.0300038158893585e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.0532868869315453,
      "train/loss": -2.0300038158893585e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 47,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010854661464691162,
      "train/pg_clipfrac": 0.0018897249829024076,
      "train/ppo_kl": 0.00022421246103476733,
      "train/entropy_loss": 0.1697261929512024
    },
    {
      "update": 47,
      "train_step": 48,
      "step": 24,
      "round": 1,
      "train/pg_loss": 0,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.251876185964631,
      "train/loss": 0,
      "train/lr-pg_1": 1e-06,
      "train/step": 48,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.00992291234433651,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.16108012199401855
    },
    {
      "update": 48,
      "train_step": 49,
      "step": 24,
      "round": 1,
      "train/pg_loss": 9.949458763003349e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.47349816994278154,
      "train/loss": 9.949458763003349e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 49,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010176289826631546,
      "train/pg_clipfrac": 0.0003204172826372087,
      "train/ppo_kl": 0.0003799013211391866,
      "train/entropy_loss": 0.15916021168231964
    },
    {
      "update": 49,
      "train_step": 50,
      "step": 25,
      "round": 1,
      "train/pg_loss": 5.587935447692871e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.7169671956051757,
      "train/loss": 5.587935447692871e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 50,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.01080232672393322,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1727149486541748
    },
    {
      "update": 50,
      "train_step": 51,
      "step": 25,
      "round": 1,
      "train/pg_loss": -2.4139881134033203e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.9100808042764664,
      "train/loss": -2.4139881134033203e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 51,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010087152943015099,
      "train/pg_clipfrac": 0.0009770516771823168,
      "train/ppo_kl": 0.00030173882259987295,
      "train/entropy_loss": 0.16281768679618835
    },
    {
      "update": 51,
      "train_step": 52,
      "step": 26,
      "round": 1,
      "train/pg_loss": 1.257285475730896e-08,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8081565641620918,
      "train/loss": 1.257285475730896e-08,
      "train/lr-pg_1": 1e-06,
      "train/step": 52,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010184688493609428,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.16533520817756653
    },
    {
      "update": 52,
      "train_step": 53,
      "step": 26,
      "round": 1,
      "train/pg_loss": 8.599832653999329e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.570711309960463,
      "train/loss": 8.599832653999329e-06,
      "train/lr-pg_1": 1e-06,
      "train/step": 53,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009580806829035282,
      "train/pg_clipfrac": 0.0006403544684872031,
      "train/ppo_kl": 0.000387724518077448,
      "train/entropy_loss": 0.15339478850364685
    },
    {
      "update": 53,
      "train_step": 54,
      "step": 27,
      "round": 1,
      "train/pg_loss": -5.820766091346741e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.2839372413277348,
      "train/loss": -5.820766091346741e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 54,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009798221290111542,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15973100066184998
    },
    {
      "update": 54,
      "train_step": 55,
      "step": 27,
      "round": 1,
      "train/pg_loss": -4.231929779052734e-06,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.6695074907358148,
      "train/loss": -4.231929779052734e-06,
      "train/lr-pg_1": 1e-06,
      "train/step": 55,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009458121843636036,
      "train/pg_clipfrac": 0.00034870341187343,
      "train/ppo_kl": 5.5183161748573184e-05,
      "train/entropy_loss": 0.15371569991111755
    },
    {
      "update": 55,
      "train_step": 56,
      "step": 28,
      "round": 1,
      "train/pg_loss": -1.862645149230957e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.483742107187182,
      "train/loss": -1.862645149230957e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 56,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.01028585247695446,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15887144207954407
    },
    {
      "update": 56,
      "train_step": 57,
      "step": 28,
      "round": 1,
      "train/pg_loss": 3.2852403819561005e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.8537527492585046,
      "train/loss": 3.2852403819561005e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 57,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009395649656653404,
      "train/pg_clipfrac": 0.0007924609817564487,
      "train/ppo_kl": 0.00046279013622552156,
      "train/entropy_loss": 0.1527433693408966
    },
    {
      "update": 57,
      "train_step": 58,
      "step": 29,
      "round": 1,
      "train/pg_loss": 1.862645149230957e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.325473157503696,
      "train/loss": 1.862645149230957e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 58,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009084174409508705,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1478612720966339
    },
    {
      "update": 58,
      "train_step": 59,
      "step": 29,
      "round": 1,
      "train/pg_loss": 0.00017364230006933212,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.356200534604092,
      "train/loss": 0.00017364230006933212,
      "train/lr-pg_1": 1e-06,
      "train/step": 59,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009799918159842491,
      "train/pg_clipfrac": 0.00210346607491374,
      "train/ppo_kl": 0.0003141792258247733,
      "train/entropy_loss": 0.15462064743041992
    },
    {
      "update": 59,
      "train_step": 60,
      "step": 30,
      "round": 1,
      "train/pg_loss": -3.725290298461914e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5821014765470276,
      "train/loss": -3.725290298461914e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 60,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009737120009958744,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15810513496398926
    },
    {
      "update": 60,
      "train_step": 61,
      "step": 30,
      "round": 1,
      "train/pg_loss": -2.813618630170822e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.9565754953552851,
      "train/loss": -2.813618630170822e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 61,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009924452751874924,
      "train/pg_clipfrac": 0.0009989954996854067,
      "train/ppo_kl": 2.7984229745925404e-05,
      "train/entropy_loss": 0.156838059425354
    },
    {
      "update": 61,
      "train_step": 62,
      "step": 31,
      "round": 1,
      "train/pg_loss": 0.00372528238222003,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.5266188830736716,
      "train/loss": 0.00372528238222003,
      "train/lr-pg_1": 1e-06,
      "train/step": 62,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009923111647367477,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1595216989517212
    },
    {
      "update": 62,
      "train_step": 63,
      "step": 31,
      "round": 1,
      "train/pg_loss": 3.683124668896198e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.8984841044404511,
      "train/loss": 3.683124668896198e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 63,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.012374823912978172,
      "train/pg_clipfrac": 0.0009348350577056408,
      "train/ppo_kl": 0.00027583728660829365,
      "train/entropy_loss": 0.1667342334985733
    },
    {
      "update": 63,
      "train_step": 64,
      "step": 32,
      "round": 1,
      "train/pg_loss": 3.725290298461914e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.4281768321975674,
      "train/loss": 3.725290298461914e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 64,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.013035958632826805,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1641625612974167
    },
    {
      "update": 64,
      "train_step": 65,
      "step": 32,
      "round": 1,
      "train/pg_loss": 0.0012595783919095993,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 34.751953070115356,
      "train/loss": 0.0012595783919095993,
      "train/lr-pg_1": 1e-06,
      "train/step": 65,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009742623195052147,
      "train/pg_clipfrac": 0.001864582416601479,
      "train/ppo_kl": 0.00040190567960962653,
      "train/entropy_loss": 0.1540728211402893
    },
    {
      "update": 65,
      "train_step": 66,
      "step": 33,
      "round": 1,
      "train/pg_loss": -2.9802322387695312e-08,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.188162131235536,
      "train/loss": -2.9802322387695312e-08,
      "train/lr-pg_1": 1e-06,
      "train/step": 66,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010145463980734348,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15846246480941772
    },
    {
      "update": 66,
      "train_step": 67,
      "step": 33,
      "round": 1,
      "train/pg_loss": 0,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0,
      "train/loss": 0,
      "train/lr-pg_1": 1e-06,
      "train/step": 67,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009777434170246124,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0.00012601655907928944,
      "train/entropy_loss": 0.15190064907073975
    },
    {
      "update": 67,
      "train_step": 68,
      "step": 34,
      "round": 1,
      "train/pg_loss": -1.0244548320770264e-08,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.3435306700760106,
      "train/loss": -1.0244548320770264e-08,
      "train/lr-pg_1": 1e-06,
      "train/step": 68,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010371606796979904,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15989571809768677
    },
    {
      "update": 68,
      "train_step": 69,
      "step": 34,
      "round": 1,
      "train/pg_loss": 0.00022800453007221222,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.271616329723406,
      "train/loss": 0.00022800453007221222,
      "train/lr-pg_1": 1e-06,
      "train/step": 69,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.01007484644651413,
      "train/pg_clipfrac": 0.0018069121288135648,
      "train/ppo_kl": 0.0004542998212855309,
      "train/entropy_loss": 0.15792518854141235
    },
    {
      "update": 69,
      "train_step": 70,
      "step": 35,
      "round": 1,
      "train/pg_loss": -3.725290298461914e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.9988618572746748,
      "train/loss": -3.725290298461914e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 70,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.013455796986818314,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1589382141828537
    },
    {
      "update": 70,
      "train_step": 71,
      "step": 35,
      "round": 1,
      "train/pg_loss": 0,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0,
      "train/loss": 0,
      "train/lr-pg_1": 1e-06,
      "train/step": 71,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009676113724708557,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0.0003508271765895188,
      "train/entropy_loss": 0.1442345231771469
    },
    {
      "update": 71,
      "train_step": 72,
      "step": 36,
      "round": 1,
      "train/pg_loss": 9.313225746154785e-10,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.9591702708680774,
      "train/loss": 9.313225746154785e-10,
      "train/lr-pg_1": 1e-06,
      "train/step": 72,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009471289813518524,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.14735782146453857
    },
    {
      "update": 72,
      "train_step": 73,
      "step": 36,
      "round": 1,
      "train/pg_loss": 6.882846355438232e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 0.9692804361011529,
      "train/loss": 6.882846355438232e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 73,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009897678159177303,
      "train/pg_clipfrac": 0.0009221415966749191,
      "train/ppo_kl": 0.0005782937514595687,
      "train/entropy_loss": 0.15042048692703247
    },
    {
      "update": 73,
      "train_step": 74,
      "step": 37,
      "round": 1,
      "train/pg_loss": 7.450580596923828e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8738292217748445,
      "train/loss": 7.450580596923828e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 74,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.00996912270784378,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.1514473557472229
    },
    {
      "update": 74,
      "train_step": 75,
      "step": 37,
      "round": 1,
      "train/pg_loss": 0.00016058865003287792,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.3375324744444457,
      "train/loss": 0.00016058865003287792,
      "train/lr-pg_1": 1e-06,
      "train/step": 75,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.00915023684501648,
      "train/pg_clipfrac": 0.0014288641978055239,
      "train/ppo_kl": 0.000544604379683733,
      "train/entropy_loss": 0.13913556933403015
    },
    {
      "update": 75,
      "train_step": 76,
      "step": 38,
      "round": 1,
      "train/pg_loss": 7.450580596923828e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.8151598520005943,
      "train/loss": 7.450580596923828e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 76,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.00943971611559391,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.13851341605186462
    },
    {
      "update": 76,
      "train_step": 77,
      "step": 38,
      "round": 1,
      "train/pg_loss": 9.970460087060928e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.0007484734414223,
      "train/loss": 9.970460087060928e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 77,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009424859657883644,
      "train/pg_clipfrac": 0.0008978781988844275,
      "train/ppo_kl": 0.000400154065573588,
      "train/entropy_loss": 0.14073032140731812
    },
    {
      "update": 77,
      "train_step": 78,
      "step": 39,
      "round": 1,
      "train/pg_loss": 4.656612873077393e-10,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.9615244196048274,
      "train/loss": 4.656612873077393e-10,
      "train/lr-pg_1": 1e-06,
      "train/step": 78,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009689480997622013,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15040069818496704
    },
    {
      "update": 78,
      "train_step": 79,
      "step": 39,
      "round": 1,
      "train/pg_loss": 1.5147961676120758e-05,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.3564241763308957,
      "train/loss": 1.5147961676120758e-05,
      "train/lr-pg_1": 1e-06,
      "train/step": 79,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009408755227923393,
      "train/pg_clipfrac": 0.0009682580130174756,
      "train/ppo_kl": 0.00039111936348490417,
      "train/entropy_loss": 0.1435682773590088
    },
    {
      "update": 79,
      "train_step": 80,
      "step": 40,
      "round": 1,
      "train/pg_loss": 4.6566128730773926e-09,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 2.2812284703741117,
      "train/loss": 4.6566128730773926e-09,
      "train/lr-pg_1": 1e-06,
      "train/step": 80,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.010082634165883064,
      "train/pg_clipfrac": 0,
      "train/ppo_kl": 0,
      "train/entropy_loss": 0.15416330099105835
    },
    {
      "update": 80,
      "train_step": 81,
      "step": 40,
      "round": 1,
      "train/pg_loss": 0.00022472860291600227,
      "train/lr-pg_0": 1e-06,
      "train/grad_norm": 1.3827675849818502,
      "train/loss": 0.00022472860291600227,
      "train/lr-pg_1": 1e-06,
      "train/step": 81,
      "train/global_batch_size": 64,
      "train/train_rollout_logprob_abs_diff": 0.009739434346556664,
      "train/pg_clipfrac": 0.0015823774738237262,
      "train/ppo_kl": 0.0003853029338642955,
      "train/entropy_loss": 0.14645734429359436
    },
    {
      "update": 81,
      "train_step": 82,
      "step": 41,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 82,
      "train/grad_norm": 0.9545119153773252,
      "train/loss": -4.6566128730773926e-09,
      "train/train_rollout_logprob_abs_diff": 0.009363018907606602,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -4.6566128730773926e-09,
      "train/entropy_loss": 0.1420472264289856
    },
    {
      "update": 82,
      "train_step": 83,
      "step": 41,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0002801147638820112,
      "train/lr-pg_0": 1e-06,
      "train/step": 83,
      "train/grad_norm": 0,
      "train/loss": 0,
      "train/train_rollout_logprob_abs_diff": 0.009174344129860401,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.1343296766281128
    },
    {
      "update": 83,
      "train_step": 84,
      "step": 42,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 84,
      "train/grad_norm": 1.8452364215462356,
      "train/loss": -3.725290298461914e-09,
      "train/train_rollout_logprob_abs_diff": 0.009416918270289898,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -3.725290298461914e-09,
      "train/entropy_loss": 0.14060211181640625
    },
    {
      "update": 84,
      "train_step": 85,
      "step": 42,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.00034345683525316417,
      "train/lr-pg_0": 1e-06,
      "train/step": 85,
      "train/grad_norm": 1.4052521874343413,
      "train/loss": 3.967806696891785e-05,
      "train/train_rollout_logprob_abs_diff": 0.010635960847139359,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001104747410863638,
      "train/pg_loss": 3.967806696891785e-05,
      "train/entropy_loss": 0.15545807778835297
    },
    {
      "update": 85,
      "train_step": 86,
      "step": 43,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 86,
      "train/grad_norm": 2.2951789874256145,
      "train/loss": 7.450580596923828e-09,
      "train/train_rollout_logprob_abs_diff": 0.010272745043039322,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 7.450580596923828e-09,
      "train/entropy_loss": 0.16141995787620544
    },
    {
      "update": 86,
      "train_step": 87,
      "step": 43,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0005683143390342593,
      "train/lr-pg_0": 1e-06,
      "train/step": 87,
      "train/grad_norm": 0.5739901080973864,
      "train/loss": -4.015583544969559e-05,
      "train/train_rollout_logprob_abs_diff": 0.009096197783946991,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.00030625949148088694,
      "train/pg_loss": -4.015583544969559e-05,
      "train/entropy_loss": 0.13957615196704865
    },
    {
      "update": 87,
      "train_step": 88,
      "step": 44,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 88,
      "train/grad_norm": 2.3105236577909496,
      "train/loss": -1.1175870895385742e-08,
      "train/train_rollout_logprob_abs_diff": 0.008936624974012375,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.1175870895385742e-08,
      "train/entropy_loss": 0.13878808915615082
    },
    {
      "update": 88,
      "train_step": 89,
      "step": 44,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.000323978514643386,
      "train/lr-pg_0": 1e-06,
      "train/step": 89,
      "train/grad_norm": 1.6274121794320933,
      "train/loss": 0.000302804633975029,
      "train/train_rollout_logprob_abs_diff": 0.009694831445813179,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.002163644414395094,
      "train/pg_loss": 0.000302804633975029,
      "train/entropy_loss": 0.14806781709194183
    },
    {
      "update": 89,
      "train_step": 90,
      "step": 45,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 90,
      "train/grad_norm": 1.0082515972417074,
      "train/loss": -3.259629011154175e-09,
      "train/train_rollout_logprob_abs_diff": 0.009695598855614662,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -3.259629011154175e-09,
      "train/entropy_loss": 0.1438772976398468
    },
    {
      "update": 90,
      "train_step": 91,
      "step": 45,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.00038578722160309553,
      "train/lr-pg_0": 1e-06,
      "train/step": 91,
      "train/grad_norm": 1.1726559897567999,
      "train/loss": 1.4457851648330688e-05,
      "train/train_rollout_logprob_abs_diff": 0.009436285123229027,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0009495363337919116,
      "train/pg_loss": 1.4457851648330688e-05,
      "train/entropy_loss": 0.14074566960334778
    },
    {
      "update": 91,
      "train_step": 92,
      "step": 46,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 92,
      "train/grad_norm": 2.0341194425938416,
      "train/loss": -1.1175870895385742e-08,
      "train/train_rollout_logprob_abs_diff": 0.009723098017275333,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.1175870895385742e-08,
      "train/entropy_loss": 0.1496860533952713
    },
    {
      "update": 92,
      "train_step": 93,
      "step": 46,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0004037920734845102,
      "train/lr-pg_0": 1e-06,
      "train/step": 93,
      "train/grad_norm": 1.2606366125885438,
      "train/loss": 0.0001459866762161255,
      "train/train_rollout_logprob_abs_diff": 0.009305693209171295,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0013925423845648766,
      "train/pg_loss": 0.0001459866762161255,
      "train/entropy_loss": 0.14039379358291626
    },
    {
      "update": 93,
      "train_step": 94,
      "step": 47,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 94,
      "train/grad_norm": 1.10267123434473,
      "train/loss": -1.862645149230957e-09,
      "train/train_rollout_logprob_abs_diff": 0.0085282102227211,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.862645149230957e-09,
      "train/entropy_loss": 0.12792342901229858
    },
    {
      "update": 94,
      "train_step": 95,
      "step": 47,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0004764458572026342,
      "train/lr-pg_0": 1e-06,
      "train/step": 95,
      "train/grad_norm": 1.4592724455714405,
      "train/loss": -2.446852158755064e-05,
      "train/train_rollout_logprob_abs_diff": 0.008798349648714066,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001693214988335967,
      "train/pg_loss": -2.446852158755064e-05,
      "train/entropy_loss": 0.12974795699119568
    },
    {
      "update": 95,
      "train_step": 96,
      "step": 48,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 96,
      "train/grad_norm": 1.3441862684469408,
      "train/loss": -1.862645149230957e-09,
      "train/train_rollout_logprob_abs_diff": 0.009383061900734901,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.862645149230957e-09,
      "train/entropy_loss": 0.13435259461402893
    },
    {
      "update": 96,
      "train_step": 97,
      "step": 48,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.00035154185025021434,
      "train/lr-pg_0": 1e-06,
      "train/step": 97,
      "train/grad_norm": 0.3897736427092652,
      "train/loss": 2.9958784580230713e-05,
      "train/train_rollout_logprob_abs_diff": 0.008169058710336685,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.00033791078021749854,
      "train/pg_loss": 2.9958784580230713e-05,
      "train/entropy_loss": 0.1161457896232605
    },
    {
      "update": 97,
      "train_step": 98,
      "step": 49,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 98,
      "train/grad_norm": 1.7734227116846248,
      "train/loss": 1.257285475730896e-08,
      "train/train_rollout_logprob_abs_diff": 0.010233457200229168,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 1.257285475730896e-08,
      "train/entropy_loss": 0.15110641717910767
    },
    {
      "update": 98,
      "train_step": 99,
      "step": 49,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0006387561443261802,
      "train/lr-pg_0": 1e-06,
      "train/step": 99,
      "train/grad_norm": 1.150510720167817,
      "train/loss": -1.5748664736747742e-05,
      "train/train_rollout_logprob_abs_diff": 0.009449069388210773,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001225376152433455,
      "train/pg_loss": -1.5748664736747742e-05,
      "train/entropy_loss": 0.13549818098545074
    },
    {
      "update": 99,
      "train_step": 100,
      "step": 50,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 100,
      "train/grad_norm": 1.586258728869074,
      "train/loss": 1.862645149230957e-09,
      "train/train_rollout_logprob_abs_diff": 0.009453503414988518,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 1.862645149230957e-09,
      "train/entropy_loss": 0.1394108235836029
    },
    {
      "update": 100,
      "train_step": 101,
      "step": 50,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0007538844365626574,
      "train/lr-pg_0": 1e-06,
      "train/step": 101,
      "train/grad_norm": 1.193201032706775,
      "train/loss": 0.00010285712778568268,
      "train/train_rollout_logprob_abs_diff": 0.00906842015683651,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0014002092648297548,
      "train/pg_loss": 0.00010285712778568268,
      "train/entropy_loss": 0.1322135329246521
    },
    {
      "update": 101,
      "train_step": 102,
      "step": 51,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 102,
      "train/grad_norm": 2.246766946499691,
      "train/loss": 2.9802322387695312e-08,
      "train/train_rollout_logprob_abs_diff": 0.00996833574026823,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 2.9802322387695312e-08,
      "train/entropy_loss": 0.15001383423805237
    },
    {
      "update": 102,
      "train_step": 103,
      "step": 51,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.000548151379916817,
      "train/lr-pg_0": 1e-06,
      "train/step": 103,
      "train/grad_norm": 1.3977672739641682,
      "train/loss": 0.00020160526037216187,
      "train/train_rollout_logprob_abs_diff": 0.009333351626992226,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001199601567350328,
      "train/pg_loss": 0.00020160526037216187,
      "train/entropy_loss": 0.14099201560020447
    },
    {
      "update": 103,
      "train_step": 104,
      "step": 52,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 104,
      "train/grad_norm": 1.9981960267063439,
      "train/loss": -9.313225746154785e-10,
      "train/train_rollout_logprob_abs_diff": 0.010662063956260681,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -9.313225746154785e-10,
      "train/entropy_loss": 0.15717121958732605
    },
    {
      "update": 104,
      "train_step": 105,
      "step": 52,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.00038372172275558114,
      "train/lr-pg_0": 1e-06,
      "train/step": 105,
      "train/grad_norm": 1.1789255776410452,
      "train/loss": 0.0001249229535460472,
      "train/train_rollout_logprob_abs_diff": 0.009416293352842331,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0015737435314804316,
      "train/pg_loss": 0.0001249229535460472,
      "train/entropy_loss": 0.13600820302963257
    },
    {
      "update": 105,
      "train_step": 106,
      "step": 53,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 106,
      "train/grad_norm": 1.8122238080287856,
      "train/loss": -2.3515895009040833e-08,
      "train/train_rollout_logprob_abs_diff": 0.009616785682737827,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -2.3515895009040833e-08,
      "train/entropy_loss": 0.14320477843284607
    },
    {
      "update": 106,
      "train_step": 107,
      "step": 53,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.00024357107758987695,
      "train/lr-pg_0": 1e-06,
      "train/step": 107,
      "train/grad_norm": 1.0328399654551197,
      "train/loss": 3.298278898000717e-05,
      "train/train_rollout_logprob_abs_diff": 0.00936039537191391,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0017974497750401497,
      "train/pg_loss": 3.298278898000717e-05,
      "train/entropy_loss": 0.13968688249588013
    },
    {
      "update": 107,
      "train_step": 108,
      "step": 54,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 108,
      "train/grad_norm": 1.2870157349579758,
      "train/loss": -1.1175870895385742e-08,
      "train/train_rollout_logprob_abs_diff": 0.009089348837733269,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.1175870895385742e-08,
      "train/entropy_loss": 0.13606682419776917
    },
    {
      "update": 108,
      "train_step": 109,
      "step": 54,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0006886614137329161,
      "train/lr-pg_0": 1e-06,
      "train/step": 109,
      "train/grad_norm": 1.4309936560404206,
      "train/loss": 5.30024990439415e-05,
      "train/train_rollout_logprob_abs_diff": 0.009706439450383186,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0017398088239133358,
      "train/pg_loss": 5.30024990439415e-05,
      "train/entropy_loss": 0.14760294556617737
    },
    {
      "update": 109,
      "train_step": 110,
      "step": 55,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 110,
      "train/grad_norm": 1.9106117126781477,
      "train/loss": 3.4924596548080444e-09,
      "train/train_rollout_logprob_abs_diff": 0.010422482155263424,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 3.4924596548080444e-09,
      "train/entropy_loss": 0.15704482793807983
    },
    {
      "update": 110,
      "train_step": 111,
      "step": 55,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0006957148434594274,
      "train/lr-pg_0": 1e-06,
      "train/step": 111,
      "train/grad_norm": 1.0435272892891738,
      "train/loss": 0.00011843256652355194,
      "train/train_rollout_logprob_abs_diff": 0.010013869032263756,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001503069419413805,
      "train/pg_loss": 0.00011843256652355194,
      "train/entropy_loss": 0.14779405295848846
    },
    {
      "update": 111,
      "train_step": 112,
      "step": 56,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 112,
      "train/grad_norm": 1.991318337320916,
      "train/loss": -1.4901161193847656e-08,
      "train/train_rollout_logprob_abs_diff": 0.010424002073705196,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -1.4901161193847656e-08,
      "train/entropy_loss": 0.1568630188703537
    },
    {
      "update": 112,
      "train_step": 113,
      "step": 56,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0006329850293695927,
      "train/lr-pg_0": 1e-06,
      "train/step": 113,
      "train/grad_norm": 1.1055869858127616,
      "train/loss": 0.00021776370704174042,
      "train/train_rollout_logprob_abs_diff": 0.0103213582187891,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.001495294738560915,
      "train/pg_loss": 0.00021776370704174042,
      "train/entropy_loss": 0.14421141147613525
    },
    {
      "update": 113,
      "train_step": 114,
      "step": 57,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 114,
      "train/grad_norm": 2.235412250193286,
      "train/loss": -6.984919309616089e-09,
      "train/train_rollout_logprob_abs_diff": 0.010112539865076542,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -6.984919309616089e-09,
      "train/entropy_loss": 0.14144402742385864
    },
    {
      "update": 114,
      "train_step": 115,
      "step": 57,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": -0.00017540244152769446,
      "train/lr-pg_0": 1e-06,
      "train/step": 115,
      "train/grad_norm": 0.8659413637524532,
      "train/loss": 8.277129381895065e-05,
      "train/train_rollout_logprob_abs_diff": 0.008939344435930252,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0005205056513659656,
      "train/pg_loss": 8.277129381895065e-05,
      "train/entropy_loss": 0.131917804479599
    },
    {
      "update": 115,
      "train_step": 116,
      "step": 58,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 116,
      "train/grad_norm": 1.5108034186847963,
      "train/loss": 4.656612873077393e-10,
      "train/train_rollout_logprob_abs_diff": 0.009727527387440205,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 4.656612873077393e-10,
      "train/entropy_loss": 0.1467668116092682
    },
    {
      "update": 116,
      "train_step": 117,
      "step": 58,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0007979543297551572,
      "train/lr-pg_0": 1e-06,
      "train/step": 117,
      "train/grad_norm": 0,
      "train/loss": 0,
      "train/train_rollout_logprob_abs_diff": 0.00931202806532383,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.13662493228912354
    },
    {
      "update": 117,
      "train_step": 118,
      "step": 59,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 118,
      "train/grad_norm": 1.8887979806578206,
      "train/loss": -2.2351741790771484e-08,
      "train/train_rollout_logprob_abs_diff": 0.010144997388124466,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": -2.2351741790771484e-08,
      "train/entropy_loss": 0.15033021569252014
    },
    {
      "update": 118,
      "train_step": 119,
      "step": 59,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.0006353874923661351,
      "train/lr-pg_0": 1e-06,
      "train/step": 119,
      "train/grad_norm": 0.769790102415513,
      "train/loss": 9.795278310775757e-05,
      "train/train_rollout_logprob_abs_diff": 0.009075315669178963,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.000525830895639956,
      "train/pg_loss": 9.795278310775757e-05,
      "train/entropy_loss": 0.13269385695457458
    },
    {
      "update": 119,
      "train_step": 120,
      "step": 60,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0,
      "train/lr-pg_0": 1e-06,
      "train/step": 120,
      "train/grad_norm": 1.6838718547495748,
      "train/loss": 0,
      "train/train_rollout_logprob_abs_diff": 0.009895390830934048,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.1488930881023407
    },
    {
      "update": 120,
      "train_step": 121,
      "step": 60,
      "round": 2,
      "train/global_batch_size": 64,
      "train/ppo_kl": 0.001362798037007451,
      "train/lr-pg_0": 1e-06,
      "train/step": 121,
      "train/grad_norm": 1.4237628183968412,
      "train/loss": 7.421150803565979e-05,
      "train/train_rollout_logprob_abs_diff": 0.00970856286585331,
      "train/lr-pg_1": 1e-06,
      "train/pg_clipfrac": 0.0016094893217086792,
      "train/pg_loss": 7.421150803565979e-05,
      "train/entropy_loss": 0.1380816549062729
    },
    {
      "update": 121,
      "train_step": 122,
      "step": 61,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 122,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -1.5366822481155396e-08,
      "train/grad_norm": 1.792100683768507,
      "train/pg_loss": -1.5366822481155396e-08,
      "train/entropy_loss": 0.15038911998271942,
      "train/train_rollout_logprob_abs_diff": 0.010026521049439907,
      "train/pg_clipfrac": 0
    },
    {
      "update": 122,
      "train_step": 123,
      "step": 61,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.000649123452603817,
      "train/step": 123,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.00015748292207717896,
      "train/grad_norm": 1.0964511215337127,
      "train/pg_loss": 0.00015748292207717896,
      "train/entropy_loss": 0.13768750429153442,
      "train/train_rollout_logprob_abs_diff": 0.009480763226747513,
      "train/pg_clipfrac": 0.0013249566545709968
    },
    {
      "update": 123,
      "train_step": 124,
      "step": 62,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 124,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 2.421438694000244e-08,
      "train/grad_norm": 2.257958430553779,
      "train/pg_loss": 2.421438694000244e-08,
      "train/entropy_loss": 0.15034103393554688,
      "train/train_rollout_logprob_abs_diff": 0.009846264496445656,
      "train/pg_clipfrac": 0
    },
    {
      "update": 124,
      "train_step": 125,
      "step": 62,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0005424279370345175,
      "train/step": 125,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 5.193520337343216e-05,
      "train/grad_norm": 0.5532536308593479,
      "train/pg_loss": 5.193520337343216e-05,
      "train/entropy_loss": 0.1497671902179718,
      "train/train_rollout_logprob_abs_diff": 0.010158082470297813,
      "train/pg_clipfrac": 0.0005546469474211335
    },
    {
      "update": 125,
      "train_step": 126,
      "step": 63,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 126,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -1.30385160446167e-08,
      "train/grad_norm": 1.4176305782718628,
      "train/pg_loss": -1.30385160446167e-08,
      "train/entropy_loss": 0.14821311831474304,
      "train/train_rollout_logprob_abs_diff": 0.009725617244839668,
      "train/pg_clipfrac": 0
    },
    {
      "update": 126,
      "train_step": 127,
      "step": 63,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.00038713778485544026,
      "train/step": 127,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 2.8546899557113647e-05,
      "train/grad_norm": 0.6964049597290219,
      "train/pg_loss": 2.8546899557113647e-05,
      "train/entropy_loss": 0.14270013570785522,
      "train/train_rollout_logprob_abs_diff": 0.00965704582631588,
      "train/pg_clipfrac": 0.000596756290178746
    },
    {
      "update": 127,
      "train_step": 128,
      "step": 64,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 128,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 1.9791092579864324,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.1499735563993454,
      "train/train_rollout_logprob_abs_diff": 0.01005636714398861,
      "train/pg_clipfrac": 0
    },
    {
      "update": 128,
      "train_step": 129,
      "step": 64,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0007333094836212695,
      "train/step": 129,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.00025458214804530144,
      "train/grad_norm": 1.2332096151640914,
      "train/pg_loss": 0.00025458214804530144,
      "train/entropy_loss": 0.1572485864162445,
      "train/train_rollout_logprob_abs_diff": 0.010701329447329044,
      "train/pg_clipfrac": 0.001916407374665141
    },
    {
      "update": 129,
      "train_step": 130,
      "step": 65,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 130,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 3.725290298461914e-09,
      "train/grad_norm": 1.562312916045964,
      "train/pg_loss": 3.725290298461914e-09,
      "train/entropy_loss": 0.1429470181465149,
      "train/train_rollout_logprob_abs_diff": 0.009518608450889587,
      "train/pg_clipfrac": 0
    },
    {
      "update": 130,
      "train_step": 131,
      "step": 65,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0002958755649160594,
      "train/step": 131,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.0001369924284517765,
      "train/grad_norm": 0.8169076283482922,
      "train/pg_loss": 0.0001369924284517765,
      "train/entropy_loss": 0.13928523659706116,
      "train/train_rollout_logprob_abs_diff": 0.009409336373209953,
      "train/pg_clipfrac": 0.0005864978302270174
    },
    {
      "update": 131,
      "train_step": 132,
      "step": 66,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 132,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -5.122274160385132e-09,
      "train/grad_norm": 1.4979675351535164,
      "train/pg_loss": -5.122274160385132e-09,
      "train/entropy_loss": 0.1364341676235199,
      "train/train_rollout_logprob_abs_diff": 0.010131368413567543,
      "train/pg_clipfrac": 0
    },
    {
      "update": 132,
      "train_step": 133,
      "step": 66,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0007741419831290841,
      "train/step": 133,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 3.29331960529089e-05,
      "train/grad_norm": 0.6685790235551751,
      "train/pg_loss": 3.29331960529089e-05,
      "train/entropy_loss": 0.13413818180561066,
      "train/train_rollout_logprob_abs_diff": 0.00916551798582077,
      "train/pg_clipfrac": 0.00046918937005102634
    },
    {
      "update": 133,
      "train_step": 134,
      "step": 67,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 134,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -3.026798367500305e-09,
      "train/grad_norm": 1.983305996084009,
      "train/pg_loss": -3.026798367500305e-09,
      "train/entropy_loss": 0.14482316374778748,
      "train/train_rollout_logprob_abs_diff": 0.00982305221259594,
      "train/pg_clipfrac": 0
    },
    {
      "update": 134,
      "train_step": 135,
      "step": 67,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.000810854253359139,
      "train/step": 135,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 1.816079020500183e-05,
      "train/grad_norm": 0.8128108017026295,
      "train/pg_loss": 1.816079020500183e-05,
      "train/entropy_loss": 0.13155588507652283,
      "train/train_rollout_logprob_abs_diff": 0.008972076699137688,
      "train/pg_clipfrac": 0.0008740989724174142
    },
    {
      "update": 135,
      "train_step": 136,
      "step": 68,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 136,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -7.450580596923828e-09,
      "train/grad_norm": 1.9122035127091819,
      "train/pg_loss": -7.450580596923828e-09,
      "train/entropy_loss": 0.13991940021514893,
      "train/train_rollout_logprob_abs_diff": 0.011473983526229858,
      "train/pg_clipfrac": 0
    },
    {
      "update": 136,
      "train_step": 137,
      "step": 68,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0005167818162590265,
      "train/step": 137,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.00019841361790895462,
      "train/grad_norm": 1.3525575271613037,
      "train/pg_loss": 0.00019841361790895462,
      "train/entropy_loss": 0.12534072995185852,
      "train/train_rollout_logprob_abs_diff": 0.008942208252847195,
      "train/pg_clipfrac": 0.001352666295133531
    },
    {
      "update": 137,
      "train_step": 138,
      "step": 69,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 138,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 2.491287887096405e-08,
      "train/grad_norm": 2.0755504062002395,
      "train/pg_loss": 2.491287887096405e-08,
      "train/entropy_loss": 0.1432851254940033,
      "train/train_rollout_logprob_abs_diff": 0.01011388748884201,
      "train/pg_clipfrac": 0
    },
    {
      "update": 138,
      "train_step": 139,
      "step": 69,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0004176949441898614,
      "train/step": 139,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 5.577458068728447e-06,
      "train/grad_norm": 0.4520008019149047,
      "train/pg_loss": 5.577458068728447e-06,
      "train/entropy_loss": 0.13757944107055664,
      "train/train_rollout_logprob_abs_diff": 0.009775247424840927,
      "train/pg_clipfrac": 0.0004935048636980355
    },
    {
      "update": 139,
      "train_step": 140,
      "step": 70,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 140,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -3.725290298461914e-09,
      "train/grad_norm": 1.170672142680387,
      "train/pg_loss": -3.725290298461914e-09,
      "train/entropy_loss": 0.14845095574855804,
      "train/train_rollout_logprob_abs_diff": 0.010181635618209839,
      "train/pg_clipfrac": 0
    },
    {
      "update": 140,
      "train_step": 141,
      "step": 70,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.000178114278241992,
      "train/step": 141,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.12329675257205963,
      "train/train_rollout_logprob_abs_diff": 0.00876616407185793,
      "train/pg_clipfrac": 0
    },
    {
      "update": 141,
      "train_step": 142,
      "step": 71,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 142,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 6.51925802230835e-09,
      "train/grad_norm": 1.732775388686583,
      "train/pg_loss": 6.51925802230835e-09,
      "train/entropy_loss": 0.1389560103416443,
      "train/train_rollout_logprob_abs_diff": 0.01081663928925991,
      "train/pg_clipfrac": 0
    },
    {
      "update": 142,
      "train_step": 143,
      "step": 71,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.00030042510479688644,
      "train/step": 143,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.12283757328987122,
      "train/train_rollout_logprob_abs_diff": 0.00910862348973751,
      "train/pg_clipfrac": 0
    },
    {
      "update": 143,
      "train_step": 144,
      "step": 72,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 144,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 9.313225746154785e-09,
      "train/grad_norm": 1.5093087791938196,
      "train/pg_loss": 9.313225746154785e-09,
      "train/entropy_loss": 0.13991384208202362,
      "train/train_rollout_logprob_abs_diff": 0.00972087774425745,
      "train/pg_clipfrac": 0
    },
    {
      "update": 144,
      "train_step": 145,
      "step": 72,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0003147454699501395,
      "train/step": 145,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.0002323593944311142,
      "train/grad_norm": 1.1872878638065658,
      "train/pg_loss": 0.0002323593944311142,
      "train/entropy_loss": 0.13151919841766357,
      "train/train_rollout_logprob_abs_diff": 0.009705128148198128,
      "train/pg_clipfrac": 0.0016039598267525434
    },
    {
      "update": 145,
      "train_step": 146,
      "step": 73,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 146,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -3.725290298461914e-09,
      "train/grad_norm": 0.8492595785621025,
      "train/pg_loss": -3.725290298461914e-09,
      "train/entropy_loss": 0.12592913210391998,
      "train/train_rollout_logprob_abs_diff": 0.010271193459630013,
      "train/pg_clipfrac": 0
    },
    {
      "update": 146,
      "train_step": 147,
      "step": 73,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.00012038215936627239,
      "train/step": 147,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.00018411222845315933,
      "train/grad_norm": 0.8151503032915643,
      "train/pg_loss": 0.00018411222845315933,
      "train/entropy_loss": 0.12292076647281647,
      "train/train_rollout_logprob_abs_diff": 0.009928662329912186,
      "train/pg_clipfrac": 0.0010928893461823463
    },
    {
      "update": 147,
      "train_step": 148,
      "step": 74,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 148,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 1.3390402205582341,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.11814321577548981,
      "train/train_rollout_logprob_abs_diff": 0.008833261206746101,
      "train/pg_clipfrac": 0
    },
    {
      "update": 148,
      "train_step": 149,
      "step": 74,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0006472551613114774,
      "train/step": 149,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.0001436220481991768,
      "train/grad_norm": 0.9220058219026634,
      "train/pg_loss": 0.0001436220481991768,
      "train/entropy_loss": 0.12434238195419312,
      "train/train_rollout_logprob_abs_diff": 0.00934553425759077,
      "train/pg_clipfrac": 0.0006558028981089592
    },
    {
      "update": 149,
      "train_step": 150,
      "step": 75,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 150,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -8.09086486697197e-09,
      "train/grad_norm": 1.7048029202304211,
      "train/pg_loss": -8.09086486697197e-09,
      "train/entropy_loss": 0.1277005672454834,
      "train/train_rollout_logprob_abs_diff": 0.009505984373390675,
      "train/pg_clipfrac": 0
    },
    {
      "update": 150,
      "train_step": 151,
      "step": 75,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.00016251765191555023,
      "train/step": 151,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 1.6796402633190155e-05,
      "train/grad_norm": 0.3855506711268579,
      "train/pg_loss": 1.6796402633190155e-05,
      "train/entropy_loss": 0.12153136730194092,
      "train/train_rollout_logprob_abs_diff": 0.01118997111916542,
      "train/pg_clipfrac": 0.0005575763061642647
    },
    {
      "update": 151,
      "train_step": 152,
      "step": 76,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 152,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 1.1175870895385742e-08,
      "train/grad_norm": 1.2643718873691383,
      "train/pg_loss": 1.1175870895385742e-08,
      "train/entropy_loss": 0.1287226676940918,
      "train/train_rollout_logprob_abs_diff": 0.010701381601393223,
      "train/pg_clipfrac": 0
    },
    {
      "update": 152,
      "train_step": 153,
      "step": 76,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.00035600888077169657,
      "train/step": 153,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 8.311308920383453e-05,
      "train/grad_norm": 1.4900152562787676,
      "train/pg_loss": 8.311308920383453e-05,
      "train/entropy_loss": 0.12307418882846832,
      "train/train_rollout_logprob_abs_diff": 0.009866556152701378,
      "train/pg_clipfrac": 0.002179433824494481
    },
    {
      "update": 153,
      "train_step": 154,
      "step": 77,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 154,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 1.3969838619232178e-09,
      "train/grad_norm": 0.9846425071419983,
      "train/pg_loss": 1.3969838619232178e-09,
      "train/entropy_loss": 0.12629345059394836,
      "train/train_rollout_logprob_abs_diff": 0.00971332285553217,
      "train/pg_clipfrac": 0
    },
    {
      "update": 154,
      "train_step": 155,
      "step": 77,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0003449422656558454,
      "train/step": 155,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 8.367933332920074e-07,
      "train/grad_norm": 0.5057180494454364,
      "train/pg_loss": 8.367933332920074e-07,
      "train/entropy_loss": 0.12692856788635254,
      "train/train_rollout_logprob_abs_diff": 0.009581619873642921,
      "train/pg_clipfrac": 0.000424699392169714
    },
    {
      "update": 155,
      "train_step": 156,
      "step": 78,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 156,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 3.725290298461914e-09,
      "train/grad_norm": 1.0715840158780492,
      "train/pg_loss": 3.725290298461914e-09,
      "train/entropy_loss": 0.1343487650156021,
      "train/train_rollout_logprob_abs_diff": 0.010201576165854931,
      "train/pg_clipfrac": 0
    },
    {
      "update": 156,
      "train_step": 157,
      "step": 78,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0011715663131326437,
      "train/step": 157,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.11357033252716064,
      "train/train_rollout_logprob_abs_diff": 0.008760424330830574,
      "train/pg_clipfrac": 0
    },
    {
      "update": 157,
      "train_step": 158,
      "step": 79,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 158,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": -3.725290298461914e-09,
      "train/grad_norm": 1.22649776992564,
      "train/pg_loss": -3.725290298461914e-09,
      "train/entropy_loss": 0.1099126785993576,
      "train/train_rollout_logprob_abs_diff": 0.00851275585591793,
      "train/pg_clipfrac": 0
    },
    {
      "update": 158,
      "train_step": 159,
      "step": 79,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0007755023543722928,
      "train/step": 159,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0.00012427940964698792,
      "train/grad_norm": 1.0099157815700563,
      "train/pg_loss": 0.00012427940964698792,
      "train/entropy_loss": 0.10083141922950745,
      "train/train_rollout_logprob_abs_diff": 0.008230337873101234,
      "train/pg_clipfrac": 0.0005200457526370883
    },
    {
      "update": 159,
      "train_step": 160,
      "step": 80,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0,
      "train/step": 160,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 3.725290298461914e-09,
      "train/grad_norm": 2.172035485794441,
      "train/pg_loss": 3.725290298461914e-09,
      "train/entropy_loss": 0.11607421934604645,
      "train/train_rollout_logprob_abs_diff": 0.009343361482024193,
      "train/pg_clipfrac": 0
    },
    {
      "update": 160,
      "train_step": 161,
      "step": 80,
      "round": 3,
      "train/lr-pg_0": 1e-06,
      "train/ppo_kl": 0.0008921445114538074,
      "train/step": 161,
      "train/global_batch_size": 64,
      "train/lr-pg_1": 1e-06,
      "train/loss": 0,
      "train/grad_norm": 0,
      "train/pg_loss": 0,
      "train/entropy_loss": 0.11616188287734985,
      "train/train_rollout_logprob_abs_diff": 0.009525912813842297,
      "train/pg_clipfrac": 0
    }
  ],
  "rounds": [
    {
      "round": 0,
      "metrics": {
        "train/entropy_loss": {
          "n": 40,
          "mean": 0.18024728074669838,
          "median": 0.18004587292671204,
          "min": 0.14106778800487518,
          "max": 0.21762660145759583
        },
        "train/ppo_kl": {
          "n": 40,
          "mean": 0.00030597948043578074,
          "median": 0.0,
          "min": -4.820346657652408e-05,
          "max": 0.0031463007908314466
        },
        "train/pg_clipfrac": {
          "n": 40,
          "mean": 0.0012273532236577011,
          "median": 0.0004293469537515193,
          "min": 0,
          "max": 0.0041798921301960945
        },
        "train/grad_norm": {
          "n": 40,
          "mean": 2.183156442757293,
          "median": 1.8257566262945866,
          "min": 1.1896283000368373,
          "max": 9.967743921483805
        },
        "train/loss": {
          "n": 40,
          "mean": 0.00012806800659745932,
          "median": 8.847564458847046e-09,
          "min": -0.007292289286851883,
          "max": 0.00592871755361557
        },
        "train/train_rollout_logprob_abs_diff": {
          "n": 40,
          "mean": 0.010232505667954684,
          "median": 0.01019962877035141,
          "min": 0.008858976885676384,
          "max": 0.011645268648862839
        },
        "rollout/response_len/mean": {
          "n": 20,
          "mean": 287.57519545941415,
          "median": 286.72036673821333,
          "min": 240.18682795698925,
          "max": 341.8862433862434
        },
        "rollout/truncated_ratio": {
          "n": 20,
          "mean": 0.11839183011434382,
          "median": 0.09673124642107273,
          "min": 0,
          "max": 0.3668032786885246
        },
        "perf/step_time": {
          "n": 20,
          "mean": 752.4373477697372,
          "median": 707.0315957069397,
          "min": 455.2762088775635,
          "max": 1205.8109483718872
        },
        "perf/rollout_time": {
          "n": 20,
          "mean": 233.969408929348,
          "median": 230.0141122341156,
          "min": 161.26529216766357,
          "max": 310.1028118133545
        },
        "perf/actor_train_tok_per_s": {
          "n": 20,
          "mean": 13538.132809438743,
          "median": 13509.752268912467,
          "min": 13251.109144790933,
          "max": 13941.810343099945
        }
      },
      "zero_grad_updates": []
    },
    {
      "round": 1,
      "metrics": {
        "train/entropy_loss": {
          "n": 40,
          "mean": 0.15556651838123797,
          "median": 0.15603913366794586,
          "min": 0.13851341605186462,
          "max": 0.1727149486541748
        },
        "train/ppo_kl": {
          "n": 40,
          "mean": 0.00022542584652001095,
          "median": 1.3992114872962702e-05,
          "min": 0,
          "max": 0.002770230872556567
        },
        "train/pg_clipfrac": {
          "n": 40,
          "mean": 0.0005807911365991458,
          "median": 0.0,
          "min": 0,
          "max": 0.0027661737985908985
        },
        "train/grad_norm": {
          "n": 40,
          "mean": 2.159832875314827,
          "median": 1.346866504524269,
          "min": 0,
          "max": 34.751953070115356
        },
        "train/loss": {
          "n": 40,
          "mean": 0.00016920817724894733,
          "median": 5.122274160385132e-09,
          "min": -2.813618630170822e-05,
          "max": 0.00372528238222003
        },
        "train/train_rollout_logprob_abs_diff": {
          "n": 40,
          "mean": 0.01015056623145938,
          "median": 0.009910295251756907,
          "min": 0.009084174409508705,
          "max": 0.013455796986818314
        },
        "rollout/response_len/mean": {
          "n": 20,
          "mean": 269.0786853690493,
          "median": 268.9059063370374,
          "min": 221.2268456375839,
          "max": 304.6536643026005
        },
        "rollout/truncated_ratio": {
          "n": 20,
          "mean": 0.2042875181365137,
          "median": 0.20121155173417377,
          "min": 0,
          "max": 0.40398740818467993
        },
        "perf/step_time": {
          "n": 20,
          "mean": 916.3522753000259,
          "median": 931.1131691932678,
          "min": 511.6666555404663,
          "max": 1396.8772304058075
        },
        "perf/rollout_time": {
          "n": 20,
          "mean": 245.10770394802094,
          "median": 244.64333403110504,
          "min": 157.11527752876282,
          "max": 332.1781198978424
        },
        "perf/actor_train_tok_per_s": {
          "n": 20,
          "mean": 13502.77871806655,
          "median": 13469.57516586745,
          "min": 13271.82583820836,
          "max": 13785.36163769729
        }
      },
      "zero_grad_updates": [
        66,
        70
      ]
    },
    {
      "round": 2,
      "metrics": {
        "train/entropy_loss": {
          "n": 40,
          "mean": 0.14212007634341717,
          "median": 0.14086884260177612,
          "min": 0.1161457896232605,
          "max": 0.16141995787620544
        },
        "train/ppo_kl": {
          "n": 40,
          "mean": 0.0002584903730166843,
          "median": 0.0,
          "min": -0.00017540244152769446,
          "max": 0.001362798037007451
        },
        "train/pg_clipfrac": {
          "n": 40,
          "mean": 0.0005634558736346662,
          "median": 0.0,
          "min": 0,
          "max": 0.002163644414395094
        },
        "train/grad_norm": {
          "n": 40,
          "mean": 1.3825304746757585,
          "median": 1.4015097306992548,
          "min": 0,
          "max": 2.3105236577909496
        },
        "train/loss": {
          "n": 40,
          "mean": 3.897383867297321e-05,
          "median": 1.1641532182693481e-09,
          "min": -4.015583544969559e-05,
          "max": 0.000302804633975029
        },
        "train/train_rollout_logprob_abs_diff": {
          "n": 40,
          "mean": 0.009577121981419624,
          "median": 0.009451286401599646,
          "min": 0.008169058710336685,
          "max": 0.010662063956260681
        },
        "rollout/response_len/mean": {
          "n": 20,
          "mean": 204.70917220175951,
          "median": 204.45139549436794,
          "min": 180.01941747572815,
          "max": 236.59295499021528
        },
        "rollout/truncated_ratio": {
          "n": 20,
          "mean": 0.22288243712157393,
          "median": 0.22085955388534173,
          "min": 0,
          "max": 0.44129158512720157
        },
        "perf/step_time": {
          "n": 20,
          "mean": 910.3257517576218,
          "median": 943.881142616272,
          "min": 444.72509241104126,
          "max": 1588.5971574783325
        },
        "perf/rollout_time": {
          "n": 20,
          "mean": 223.9434867978096,
          "median": 229.04083669185638,
          "min": 137.54658913612366,
          "max": 318.23707246780396
        },
        "perf/actor_train_tok_per_s": {
          "n": 20,
          "mean": 13468.471235521616,
          "median": 13437.831894123661,
          "min": 13161.809954681665,
          "max": 13959.616697463496
        }
      },
      "zero_grad_updates": [
        82,
        116
      ]
    },
    {
      "round": 3,
      "metrics": {
        "train/entropy_loss": {
          "n": 40,
          "mean": 0.13257721327245237,
          "median": 0.13284703344106674,
          "min": 0.10083141922950745,
          "max": 0.1572485864162445
        },
        "train/ppo_kl": {
          "n": 40,
          "mean": 0.00025977378427342044,
          "median": 6.0191079683136195e-05,
          "min": 0,
          "max": 0.0011715663131326437
        },
        "train/pg_clipfrac": {
          "n": 40,
          "mean": 0.000380078298621811,
          "median": 0.0,
          "min": 0,
          "max": 0.002179433824494481
        },
        "train/grad_norm": {
          "n": 40,
          "mean": 1.1375237217546175,
          "median": 1.1789800032434763,
          "min": 0,
          "max": 2.257958430553779
        },
        "train/loss": {
          "n": 40,
          "mean": 4.174423665972427e-05,
          "median": 7.916241884231567e-09,
          "min": -1.5366822481155396e-08,
          "max": 0.00025458214804530144
        },
        "train/train_rollout_logprob_abs_diff": {
          "n": 40,
          "mean": 0.009719705046154558,
          "median": 0.00971710029989481,
          "min": 0.008230337873101234,
          "max": 0.011473983526229858
        },
        "rollout/response_len/mean": {
          "n": 20,
          "mean": 246.79088740565112,
          "median": 243.65196617537833,
          "min": 210.86363636363637,
          "max": 289.44101123595505
        },
        "rollout/truncated_ratio": {
          "n": 20,
          "mean": 0.1579718473950813,
          "median": 0.17043076198005774,
          "min": 0.00487012987012987,
          "max": 0.33052434456928836
        },
        "perf/step_time": {
          "n": 20,
          "mean": 757.6003757953644,
          "median": 776.6111532449722,
          "min": 469.6946506500244,
          "max": 1322.1463577747345
        },
        "perf/rollout_time": {
          "n": 20,
          "mean": 208.76262168884278,
          "median": 213.04073417186737,
          "min": 142.12779879570007,
          "max": 300.60658049583435
        },
        "perf/actor_train_tok_per_s": {
          "n": 20,
          "mean": 13527.106150288168,
          "median": 13480.5463031252,
          "min": 13314.823681485666,
          "max": 13956.890246986106
        }
      },
      "zero_grad_updates": [
        140,
        142,
        156,
        160
      ]
    }
  ],
  "reconciliation": {
    "compared_numeric_values": 3118,
    "max_abs_error": 0.0
  },
  "step_mapping": "W&B train/step is 2..161; display update = train/step \u2212 1; rollout step = floor(train/step / 2). W&B _step is a logging event index.",
  "scope": "Four completed rounds only, rollout steps 1..80, 160 optimizer updates"
}
