{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00018,
  "eval_steps": 500,
  "global_step": 9,
  "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": 3.0,
      "completions/max_terminated_length": 3.0,
      "completions/mean_length": 2.875,
      "completions/mean_terminated_length": 2.875,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "entropy": 0.056345001052250154,
      "epoch": 2e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.00689673563465476,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": 0.0,
      "num_tokens": 92675.0,
      "reward": -0.03840970993041992,
      "reward_std": 6.371873378753662,
      "rewards/rollout_reward_func/mean": -0.03840970993041992,
      "rewards/rollout_reward_func/std": 6.371872901916504,
      "sampling/importance_sampling_ratio/max": 1.1443473100662231,
      "sampling/importance_sampling_ratio/mean": 0.9763789176940918,
      "sampling/importance_sampling_ratio/min": 0.24068869650363922,
      "sampling/sampling_logp_difference/max": 1.4758713245391846,
      "sampling/sampling_logp_difference/mean": 0.022954855114221573,
      "step": 1,
      "step_time": 44.682924225000534
    },
    {
      "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": 0.056345001052250154,
      "epoch": 4e-05,
      "grad_norm": 0.006772701628506184,
      "kl": 0.0,
      "learning_rate": 2.2857142857142855e-07,
      "loss": 0.0,
      "step": 2,
      "step_time": 9.55681869000091
    },
    {
      "clip_ratio/high_max": 0.02083333395421505,
      "clip_ratio/high_mean": 0.010416666977107525,
      "clip_ratio/low_mean": 0.012500000186264515,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.02291666716337204,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3.0,
      "completions/max_terminated_length": 3.0,
      "completions/mean_length": 2.90625,
      "completions/mean_terminated_length": 2.90625,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "entropy": 0.10085208504460752,
      "epoch": 6e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.008757508359849453,
      "kl": 0.004195388915081644,
      "learning_rate": 4.571428571428571e-07,
      "loss": 0.0,
      "num_tokens": 185014.0,
      "reward": -1.9113237857818604,
      "reward_std": 3.7397446632385254,
      "rewards/rollout_reward_func/mean": -1.9113237857818604,
      "rewards/rollout_reward_func/std": 3.7397446632385254,
      "sampling/importance_sampling_ratio/max": 1.2321388721466064,
      "sampling/importance_sampling_ratio/mean": 0.9895046353340149,
      "sampling/importance_sampling_ratio/min": 0.5146191716194153,
      "sampling/sampling_logp_difference/max": 0.4437241554260254,
      "sampling/sampling_logp_difference/mean": 0.024998027831315994,
      "step": 3,
      "step_time": 44.41287102899878
    },
    {
      "clip_ratio/high_max": 0.06250000186264515,
      "clip_ratio/high_mean": 0.031250000931322575,
      "clip_ratio/low_mean": 0.010416666977107525,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0416666679084301,
      "entropy": 0.10115758149186149,
      "epoch": 8e-05,
      "grad_norm": 0.009121721610426903,
      "kl": 0.01587161054246744,
      "learning_rate": 6.857142857142857e-07,
      "loss": 0.0,
      "step": 4,
      "step_time": 10.027988500997708
    },
    {
      "clip_ratio/high_max": 0.0416666679084301,
      "clip_ratio/high_mean": 0.02083333395421505,
      "clip_ratio/low_mean": 0.026041666977107525,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.046875000931322575,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3.0,
      "completions/max_terminated_length": 3.0,
      "completions/mean_length": 2.78125,
      "completions/mean_terminated_length": 2.78125,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "entropy": 0.12059010594384745,
      "epoch": 0.0001,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.005114904139190912,
      "kl": 0.01495055931777678,
      "learning_rate": 9.142857142857142e-07,
      "loss": 0.0,
      "num_tokens": 279025.0,
      "reward": -2.9652585983276367,
      "reward_std": 4.270595550537109,
      "rewards/rollout_reward_func/mean": -2.9652585983276367,
      "rewards/rollout_reward_func/std": 4.270596027374268,
      "sampling/importance_sampling_ratio/max": 1.40328049659729,
      "sampling/importance_sampling_ratio/mean": 0.9845530390739441,
      "sampling/importance_sampling_ratio/min": 0.4980602264404297,
      "sampling/sampling_logp_difference/max": 0.7273361086845398,
      "sampling/sampling_logp_difference/mean": 0.03421199321746826,
      "step": 5,
      "step_time": 44.15886100900025
    },
    {
      "clip_ratio/high_max": 0.0416666679084301,
      "clip_ratio/high_mean": 0.02083333395421505,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.02083333395421505,
      "entropy": 0.13055883452761918,
      "epoch": 0.00012,
      "grad_norm": 0.06413441896438599,
      "kl": 0.010965288731227929,
      "learning_rate": 1.1428571428571428e-06,
      "loss": 0.0001,
      "step": 6,
      "step_time": 9.721866401000625
    },
    {
      "clip_ratio/high_max": 0.04583333432674408,
      "clip_ratio/high_mean": 0.02291666716337204,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.02291666716337204,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3.0,
      "completions/max_terminated_length": 3.0,
      "completions/mean_length": 2.78125,
      "completions/mean_terminated_length": 2.78125,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "entropy": 0.0803429854568094,
      "epoch": 0.00014,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.007241566199809313,
      "kl": 0.008746063784201397,
      "learning_rate": 1.3714285714285715e-06,
      "loss": -0.0001,
      "num_tokens": 370157.0,
      "reward": -1.6896233558654785,
      "reward_std": 8.037113189697266,
      "rewards/rollout_reward_func/mean": -1.6896233558654785,
      "rewards/rollout_reward_func/std": 8.037113189697266,
      "sampling/importance_sampling_ratio/max": 1.8139108419418335,
      "sampling/importance_sampling_ratio/mean": 0.9827349185943604,
      "sampling/importance_sampling_ratio/min": 0.28889480233192444,
      "sampling/sampling_logp_difference/max": 1.241692066192627,
      "sampling/sampling_logp_difference/mean": 0.043297506868839264,
      "step": 7,
      "step_time": 43.00307397600045
    },
    {
      "clip_ratio/high_max": 0.02083333395421505,
      "clip_ratio/high_mean": 0.010416666977107525,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.010416666977107525,
      "entropy": 0.07465657824650407,
      "epoch": 0.00016,
      "grad_norm": 0.012575524859130383,
      "kl": 0.005579028980719158,
      "learning_rate": 1.6e-06,
      "loss": -0.0001,
      "step": 8,
      "step_time": 9.592408863998571
    },
    {
      "clip_ratio/high_max": 0.02083333395421505,
      "clip_ratio/high_mean": 0.010416666977107525,
      "clip_ratio/low_mean": 0.010416666977107525,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.02083333395421505,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 3.0,
      "completions/max_terminated_length": 3.0,
      "completions/mean_length": 2.9375,
      "completions/mean_terminated_length": 2.9375,
      "completions/min_length": 2.0,
      "completions/min_terminated_length": 2.0,
      "entropy": 0.10514851659536362,
      "epoch": 0.00018,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 0.01258761901408434,
      "kl": 0.0067734302374447,
      "learning_rate": 1.8285714285714284e-06,
      "loss": 0.0,
      "num_tokens": 462142.0,
      "reward": -0.16374534368515015,
      "reward_std": 4.302643775939941,
      "rewards/rollout_reward_func/mean": -0.16374534368515015,
      "rewards/rollout_reward_func/std": 4.302643299102783,
      "sampling/importance_sampling_ratio/max": 1.2026947736740112,
      "sampling/importance_sampling_ratio/mean": 0.9182904958724976,
      "sampling/importance_sampling_ratio/min": 0.23646476864814758,
      "sampling/sampling_logp_difference/max": 1.441957950592041,
      "sampling/sampling_logp_difference/mean": 0.04875129088759422,
      "step": 9,
      "step_time": 43.266243381001914
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 100000,
  "num_input_tokens_seen": 462142,
  "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
}