{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5,
  "eval_steps": 500,
  "global_step": 8,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "buffer_qsize": 0.0,
      "clip_ratio/high_max": 3.222041777917184e-05,
      "clip_ratio/high_mean": 3.222041777917184e-05,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 3.222041777917184e-05,
      "completions/mean_length": 3876.25,
      "entropy": 0.9389231279492378,
      "epoch": 0.25,
      "forward_time_s": 0.30089330673217773,
      "generation_tok_per_s": 6915.577392578125,
      "grad_norm": 0.36549293994903564,
      "kl": 0.0003301510150777176,
      "learning_rate": 1e-06,
      "loss": 0.02711719088256359,
      "queue_wait_time_s": 2.0328194051980972,
      "ratio": 1.0000957623124123,
      "reward": 0.4375,
      "reward_std": 0.49206146597862244,
      "rewards/r1_distill_math_reward": 0.4375,
      "scoring_time_ms": 18.674388885498047,
      "step": 1,
      "step_time": 6.262407066999003,
      "train_seq_len": 4371.125,
      "training_tok/s": 12972.816542826044,
      "wait_scoring_ms": 0.20329749584197998,
      "weight_sync_time_s": 0.14123249053955078
    },
    {
      "buffer_qsize": 3.5,
      "clip_ratio/high_max": 1.6543144738534465e-05,
      "clip_ratio/high_mean": 1.6543144738534465e-05,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 1.6543144738534465e-05,
      "completions/mean_length": 6120.375,
      "entropy": 0.7296525910496712,
      "epoch": 0.5,
      "forward_time_s": 0.25667285919189453,
      "generation_tok_per_s": 13383.16943359375,
      "grad_norm": 0.2867599427700043,
      "iteration_time_s": 7.6814478730084375,
      "kl": 0.00025582036323612556,
      "learning_rate": 1e-06,
      "loss": 3.546476364135742e-05,
      "queue_wait_time_s": 0.3530673235654831,
      "ratio": 1.0000218078494072,
      "reward": 0.75,
      "reward_std": 0.4330126941204071,
      "rewards/r1_distill_math_reward": 0.75,
      "scoring_time_ms": 5.158496618270874,
      "step": 2,
      "step_time": 7.5191129069717135,
      "train_seq_len": 6374.125,
      "training_tok/s": 13318.513123161385,
      "wait_scoring_ms": 5.808052182197571
    },
    {
      "buffer_qsize": 18.0,
      "clip_ratio/high_max": 7.113443098205607e-05,
      "clip_ratio/high_mean": 7.113443098205607e-05,
      "clip_ratio/low_mean": 1.5646326573914848e-05,
      "clip_ratio/low_min": 1.5646326573914848e-05,
      "clip_ratio/region_mean": 8.678075755597092e-05,
      "completions/mean_length": 6053.8125,
      "entropy": 0.8014494031667709,
      "epoch": 0.75,
      "forward_time_s": 0.2557235360145569,
      "generation_tok_per_s": 13382.43359375,
      "grad_norm": 0.3024653494358063,
      "iteration_time_s": 7.571961241017561,
      "kl": 0.00028099300106987357,
      "learning_rate": 1e-06,
      "loss": 0.00955304503440857,
      "queue_wait_time_s": 0.00046597421169281006,
      "ratio": 1.0000539049506187,
      "reward": 0.5625,
      "reward_std": 0.38186579942703247,
      "rewards/r1_distill_math_reward": 0.5625,
      "scoring_time_ms": 0.7072950005531311,
      "step": 3,
      "step_time": 7.4889817090006545,
      "train_seq_len": 6390.375,
      "training_tok/s": 13393.731171849211,
      "wait_scoring_ms": 12.160233974456787
    },
    {
      "buffer_qsize": 34.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 4.13285370086669e-05,
      "clip_ratio/low_min": 4.13285370086669e-05,
      "clip_ratio/region_mean": 4.13285370086669e-05,
      "completions/mean_length": 6813.625,
      "entropy": 0.811307743191719,
      "epoch": 1.0,
      "forward_time_s": 0.2839837670326233,
      "generation_tok_per_s": 13381.1396484375,
      "grad_norm": 0.2735491394996643,
      "iteration_time_s": 8.428669046988944,
      "kl": 0.00027251767460256815,
      "learning_rate": 1e-06,
      "loss": 0.038605645298957825,
      "queue_wait_time_s": 0.0004654228687286377,
      "ratio": 0.9999778047204018,
      "reward": 0.25,
      "reward_std": 0.4330126941204071,
      "rewards/r1_distill_math_reward": 0.25,
      "scoring_time_ms": 1.8631829619407654,
      "step": 4,
      "step_time": 8.344294181006262,
      "train_seq_len": 7421.5,
      "training_tok/s": 13408.494449990198,
      "wait_scoring_ms": 14.634688377380371,
      "weight_sync_time_s": 0.15913748741149902
    },
    {
      "buffer_qsize": 0.0,
      "clip_ratio/high_max": 1.8846135390049312e-05,
      "clip_ratio/high_mean": 1.8846135390049312e-05,
      "clip_ratio/low_mean": 1.011736094369553e-05,
      "clip_ratio/low_min": 1.011736094369553e-05,
      "clip_ratio/region_mean": 2.8963496333744843e-05,
      "completions/mean_length": 5484.0,
      "entropy": 0.7321677580475807,
      "epoch": 0.125,
      "forward_time_s": 0.3462470471858978,
      "generation_tok_per_s": 9634.451171875,
      "grad_norm": 0.28255149722099304,
      "kl": 0.00025452028421568684,
      "learning_rate": 1e-06,
      "loss": -0.004137326031923294,
      "queue_wait_time_s": 1.8160016685724258,
      "ratio": 1.0000781491398811,
      "reward": 0.5625,
      "reward_std": 0.38186579942703247,
      "rewards/r1_distill_math_reward": 0.5625,
      "scoring_time_ms": 13.104030728340149,
      "step": 5,
      "step_time": 7.895209455047734,
      "train_seq_len": 5787.375,
      "training_tok/s": 13242.969307036337,
      "wait_scoring_ms": 0.1778009943664074,
      "weight_sync_time_s": 0.13536810874938965
    },
    {
      "buffer_qsize": 3.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/mean_length": 6134.3125,
      "entropy": 0.6220929026603699,
      "epoch": 0.25,
      "forward_time_s": 0.26168057322502136,
      "generation_tok_per_s": 11878.65380859375,
      "grad_norm": 0.17205984890460968,
      "iteration_time_s": 7.687154909974197,
      "kl": 0.00023902823340904433,
      "learning_rate": 1e-06,
      "loss": 0.0026915371417999268,
      "queue_wait_time_s": 0.3177753835916519,
      "ratio": 0.9999898672103882,
      "reward": 0.375,
      "reward_std": 0.21650634706020355,
      "rewards/r1_distill_math_reward": 0.375,
      "scoring_time_ms": 8.408703327178955,
      "step": 6,
      "step_time": 7.5453005689778365,
      "train_seq_len": 6619.25,
      "training_tok/s": 13203.215699232549,
      "wait_scoring_ms": 0.5662760138511658
    },
    {
      "buffer_qsize": 18.0,
      "clip_ratio/high_max": 5.6621894145791885e-05,
      "clip_ratio/high_mean": 5.6621894145791885e-05,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 5.6621894145791885e-05,
      "completions/mean_length": 5713.375,
      "entropy": 0.8666651099920273,
      "epoch": 0.375,
      "forward_time_s": 0.2426280975341797,
      "generation_tok_per_s": 13567.3173828125,
      "grad_norm": 0.30514439940452576,
      "iteration_time_s": 7.157755050022388,
      "kl": 0.0002972106722154422,
      "learning_rate": 1e-06,
      "loss": 0.015170708298683167,
      "queue_wait_time_s": 0.0004561096429824829,
      "ratio": 1.0000762566924095,
      "reward": 0.75,
      "reward_std": 0.40742091834545135,
      "rewards/r1_distill_math_reward": 0.75,
      "scoring_time_ms": 2.2001885175704956,
      "step": 7,
      "step_time": 7.077003714046441,
      "train_seq_len": 5968.5,
      "training_tok/s": 13234.408437249378,
      "wait_scoring_ms": 14.30916166305542
    },
    {
      "buffer_qsize": 34.0,
      "clip_ratio/high_max": 9.804690307646524e-06,
      "clip_ratio/high_mean": 9.804690307646524e-06,
      "clip_ratio/low_mean": 3.2692873901396524e-05,
      "clip_ratio/low_min": 3.2692873901396524e-05,
      "clip_ratio/region_mean": 4.249756420904305e-05,
      "completions/mean_length": 5408.0625,
      "entropy": 0.9108624383807182,
      "epoch": 0.5,
      "forward_time_s": 0.23351335525512695,
      "generation_tok_per_s": 13553.60693359375,
      "grad_norm": 0.3738342225551605,
      "iteration_time_s": 6.897600525000598,
      "kl": 0.0003173556542606093,
      "learning_rate": 1e-06,
      "loss": 0.0034758001565933228,
      "queue_wait_time_s": 0.0004056990146636963,
      "ratio": 0.9998656511306763,
      "reward": 0.5,
      "reward_std": 0.4330126941204071,
      "rewards/r1_distill_math_reward": 0.5,
      "scoring_time_ms": 18.20103907585144,
      "step": 8,
      "step_time": 6.822746234043734,
      "train_seq_len": 5848.75,
      "training_tok/s": 13207.699007613326,
      "wait_scoring_ms": 35.41268348693848,
      "weight_sync_time_s": 0.23273587226867676
    }
  ],
  "logging_steps": 1,
  "max_steps": 8,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 4,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6909415142986752.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}