{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00024,
  "eval_steps": 500,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "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/clipped_ratio": 0.0,
      "completions/max_length": 1928.0,
      "completions/max_terminated_length": 1928.0,
      "completions/mean_length": 1755.5,
      "completions/mean_terminated_length": 1755.5,
      "completions/min_length": 1363.0,
      "completions/min_terminated_length": 1363.0,
      "entropy": 0.5008892454206944,
      "epoch": 2e-05,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 1.3115488290786743,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": -0.013,
      "num_tokens": 79465.0,
      "reward": -0.06562499701976776,
      "reward_std": 0.11540064215660095,
      "rewards/rollout_reward_func/mean": -0.06562499701976776,
      "rewards/rollout_reward_func/std": 0.19278831779956818,
      "sampling/importance_sampling_ratio/max": 1.7688987255096436,
      "sampling/importance_sampling_ratio/mean": 1.0469263792037964,
      "sampling/importance_sampling_ratio/min": 0.6919283866882324,
      "sampling/sampling_logp_difference/max": 0.3354456424713135,
      "sampling/sampling_logp_difference/mean": 0.018323732540011406,
      "step": 1,
      "step_time": 37.61465689600027
    },
    {
      "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,
      "entropy": 0.5008892454206944,
      "epoch": 4e-05,
      "grad_norm": 1.3185971975326538,
      "kl": 0.0,
      "learning_rate": 2.8571428571428575e-07,
      "loss": -0.013,
      "step": 2,
      "step_time": 6.174051868999413
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.00390625,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005859375,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1859.0,
      "completions/max_terminated_length": 1859.0,
      "completions/mean_length": 1702.625,
      "completions/mean_terminated_length": 1702.625,
      "completions/min_length": 1369.0,
      "completions/min_terminated_length": 1369.0,
      "entropy": 0.5293242037296295,
      "epoch": 6e-05,
      "frac_reward_zero_std": 0.375,
      "grad_norm": 2.0585877895355225,
      "kl": 0.0012533214612631127,
      "learning_rate": 5.714285714285715e-07,
      "loss": 0.0512,
      "num_tokens": 157637.0,
      "reward": -0.06562500447034836,
      "reward_std": 0.07261751592159271,
      "rewards/rollout_reward_func/mean": -0.06562500447034836,
      "rewards/rollout_reward_func/std": 0.11807426810264587,
      "sampling/importance_sampling_ratio/max": 2.1121888160705566,
      "sampling/importance_sampling_ratio/mean": 1.0355395078659058,
      "sampling/importance_sampling_ratio/min": 0.5390263199806213,
      "sampling/sampling_logp_difference/max": 0.683211088180542,
      "sampling/sampling_logp_difference/mean": 0.02265790104866028,
      "step": 3,
      "step_time": 34.844595972002026
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0035556890070438385,
      "clip_ratio/low_min": 0.0032051282469183207,
      "clip_ratio/region_mean": 0.0055088140070438385,
      "entropy": 0.5292382091283798,
      "epoch": 8e-05,
      "grad_norm": 2.174487829208374,
      "kl": 0.0010874079816858284,
      "learning_rate": 8.571428571428572e-07,
      "loss": 0.0533,
      "step": 4,
      "step_time": 6.182152846000463
    },
    {
      "clip_ratio/high_max": 0.0032051282469183207,
      "clip_ratio/high_mean": 0.0016025641234591603,
      "clip_ratio/low_mean": 0.001953125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035556891234591603,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1825.0,
      "completions/max_terminated_length": 1825.0,
      "completions/mean_length": 1647.34375,
      "completions/mean_terminated_length": 1647.34375,
      "completions/min_length": 1082.0,
      "completions/min_terminated_length": 1082.0,
      "entropy": 0.5518431439995766,
      "epoch": 0.0001,
      "frac_reward_zero_std": 0.125,
      "grad_norm": 2.0179476737976074,
      "kl": 0.0010270353013765998,
      "learning_rate": 1.142857142857143e-06,
      "loss": 0.0455,
      "num_tokens": 233699.0,
      "reward": -0.07500000298023224,
      "reward_std": 0.1056123897433281,
      "rewards/rollout_reward_func/mean": -0.07500000298023224,
      "rewards/rollout_reward_func/std": 0.11359237134456635,
      "sampling/importance_sampling_ratio/max": 1.5899720191955566,
      "sampling/importance_sampling_ratio/mean": 1.0574636459350586,
      "sampling/importance_sampling_ratio/min": 0.7474462389945984,
      "sampling/sampling_logp_difference/max": 0.23842597007751465,
      "sampling/sampling_logp_difference/mean": 0.02176746539771557,
      "step": 5,
      "step_time": 36.2267263499989
    },
    {
      "clip_ratio/high_max": 0.0032051282469183207,
      "clip_ratio/high_mean": 0.0016025641234591603,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0016025641234591603,
      "entropy": 0.551122885197401,
      "epoch": 0.00012,
      "grad_norm": 2.1906440258026123,
      "kl": 0.0010457161733938847,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 0.0423,
      "step": 6,
      "step_time": 6.1078073650005535
    },
    {
      "clip_ratio/high_max": 0.005681818351149559,
      "clip_ratio/high_mean": 0.0028409091755747795,
      "clip_ratio/low_mean": 0.0026041667442768812,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.005445075919851661,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1803.0,
      "completions/max_terminated_length": 1803.0,
      "completions/mean_length": 1562.1875,
      "completions/mean_terminated_length": 1562.1875,
      "completions/min_length": 359.0,
      "completions/min_terminated_length": 359.0,
      "entropy": 0.48404254019260406,
      "epoch": 0.00014,
      "frac_reward_zero_std": 0.25,
      "grad_norm": 1.9169107675552368,
      "kl": 0.001029455175739713,
      "learning_rate": 1.7142857142857145e-06,
      "loss": -0.0132,
      "num_tokens": 307662.0,
      "reward": -0.06875000894069672,
      "reward_std": 0.20966878533363342,
      "rewards/rollout_reward_func/mean": -0.06875000894069672,
      "rewards/rollout_reward_func/std": 0.3440156579017639,
      "sampling/importance_sampling_ratio/max": 1.3860647678375244,
      "sampling/importance_sampling_ratio/mean": 1.0004303455352783,
      "sampling/importance_sampling_ratio/min": 0.6425995826721191,
      "sampling/sampling_logp_difference/max": 0.30062544345855713,
      "sampling/sampling_logp_difference/mean": 0.02074645832180977,
      "step": 7,
      "step_time": 35.09266174299955
    },
    {
      "clip_ratio/high_max": 0.0037878789007663727,
      "clip_ratio/high_mean": 0.0018939394503831863,
      "clip_ratio/low_mean": 0.0026041667442768812,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.004498106427490711,
      "entropy": 0.48381197452545166,
      "epoch": 0.00016,
      "grad_norm": 1.8268851041793823,
      "kl": 0.001109405180613976,
      "learning_rate": 2.0000000000000003e-06,
      "loss": -0.0145,
      "step": 8,
      "step_time": 6.09429400800218
    },
    {
      "clip_ratio/high_max": 0.0078125,
      "clip_ratio/high_mean": 0.005800189450383186,
      "clip_ratio/low_mean": 0.003791360300965607,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.009591549751348794,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1877.0,
      "completions/max_terminated_length": 1877.0,
      "completions/mean_length": 1641.03125,
      "completions/mean_terminated_length": 1641.03125,
      "completions/min_length": 294.0,
      "completions/min_terminated_length": 294.0,
      "entropy": 0.5420897267758846,
      "epoch": 0.00018,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 2.5005621910095215,
      "kl": 0.0013470195408444852,
      "learning_rate": 2.285714285714286e-06,
      "loss": -0.0267,
      "num_tokens": 383747.0,
      "reward": -0.10625000298023224,
      "reward_std": 0.16220411658287048,
      "rewards/rollout_reward_func/mean": -0.10625000298023224,
      "rewards/rollout_reward_func/std": 0.19499793648719788,
      "sampling/importance_sampling_ratio/max": 1.596044898033142,
      "sampling/importance_sampling_ratio/mean": 0.9659844636917114,
      "sampling/importance_sampling_ratio/min": 0.42962342500686646,
      "sampling/sampling_logp_difference/max": 0.3873096704483032,
      "sampling/sampling_logp_difference/mean": 0.02420671284198761,
      "step": 9,
      "step_time": 35.05370295100056
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.001953125,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00390625,
      "entropy": 0.5415336266160011,
      "epoch": 0.0002,
      "grad_norm": 2.5184950828552246,
      "kl": 0.0014528355168295093,
      "learning_rate": 2.571428571428571e-06,
      "loss": -0.0249,
      "step": 10,
      "step_time": 6.594039571999929
    },
    {
      "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/clipped_ratio": 0.0,
      "completions/max_length": 1835.0,
      "completions/max_terminated_length": 1835.0,
      "completions/mean_length": 1630.1875,
      "completions/mean_terminated_length": 1630.1875,
      "completions/min_length": 1176.0,
      "completions/min_terminated_length": 1176.0,
      "entropy": 0.5399914123117924,
      "epoch": 0.00022,
      "frac_reward_zero_std": 0.125,
      "grad_norm": 1.7063910961151123,
      "kl": 0.0010663218854460865,
      "learning_rate": 2.8571428571428573e-06,
      "loss": -0.0121,
      "num_tokens": 459548.0,
      "reward": -0.08124999701976776,
      "reward_std": 0.07863743603229523,
      "rewards/rollout_reward_func/mean": -0.08124999701976776,
      "rewards/rollout_reward_func/std": 0.11198358237743378,
      "sampling/importance_sampling_ratio/max": 1.3067278861999512,
      "sampling/importance_sampling_ratio/mean": 0.980239748954773,
      "sampling/importance_sampling_ratio/min": 0.6372140645980835,
      "sampling/sampling_logp_difference/max": 0.36971473693847656,
      "sampling/sampling_logp_difference/mean": 0.021189481019973755,
      "step": 11,
      "step_time": 34.695202157001404
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.001953125,
      "entropy": 0.5411706231534481,
      "epoch": 0.00024,
      "grad_norm": 1.7831242084503174,
      "kl": 0.0010139764053747058,
      "learning_rate": 3.142857142857143e-06,
      "loss": -0.0092,
      "step": 12,
      "step_time": 6.109187065001606
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 100000,
  "num_input_tokens_seen": 459548,
  "num_train_epochs": 2,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}