{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00022,
  "eval_steps": 500,
  "global_step": 11,
  "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": 1838.0,
      "completions/max_terminated_length": 1838.0,
      "completions/mean_length": 1617.0625,
      "completions/mean_terminated_length": 1617.0625,
      "completions/min_length": 1177.0,
      "completions/min_terminated_length": 1177.0,
      "entropy": 0.5484257489442825,
      "epoch": 2e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.9166723489761353,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": -0.0845,
      "num_tokens": 75257.0,
      "reward": -0.5254741907119751,
      "reward_std": 0.1556793451309204,
      "rewards/rollout_reward_func/mean": -0.5254741907119751,
      "rewards/rollout_reward_func/std": 0.19436030089855194,
      "sampling/importance_sampling_ratio/max": 1.6091933250427246,
      "sampling/importance_sampling_ratio/mean": 1.009643793106079,
      "sampling/importance_sampling_ratio/min": 0.7761501669883728,
      "sampling/sampling_logp_difference/max": 0.2959944009780884,
      "sampling/sampling_logp_difference/mean": 0.02158474177122116,
      "step": 1,
      "step_time": 39.089814388993545
    },
    {
      "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": 0.5484257489442825,
      "epoch": 4e-05,
      "grad_norm": 1.95936918258667,
      "kl": 0.0,
      "learning_rate": 2.8571428571428575e-07,
      "loss": -0.0845,
      "step": 2,
      "step_time": 6.1725359009979
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0016025641234591603,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035556891234591603,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1872.0,
      "completions/max_terminated_length": 1872.0,
      "completions/mean_length": 1648.46875,
      "completions/mean_terminated_length": 1648.46875,
      "completions/min_length": 1152.0,
      "completions/min_terminated_length": 1152.0,
      "entropy": 0.5180552825331688,
      "epoch": 6e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 2.2347609996795654,
      "kl": 0.0011785991373471916,
      "learning_rate": 5.714285714285715e-07,
      "loss": 0.1137,
      "num_tokens": 151661.0,
      "reward": -0.6038702726364136,
      "reward_std": 0.22588981688022614,
      "rewards/rollout_reward_func/mean": -0.6038702726364136,
      "rewards/rollout_reward_func/std": 0.2508644163608551,
      "sampling/importance_sampling_ratio/max": 1.4536101818084717,
      "sampling/importance_sampling_ratio/mean": 1.0416083335876465,
      "sampling/importance_sampling_ratio/min": 0.705021858215332,
      "sampling/sampling_logp_difference/max": 0.31142354011535645,
      "sampling/sampling_logp_difference/mean": 0.02324344590306282,
      "step": 3,
      "step_time": 33.94999921999988
    },
    {
      "clip_ratio/high_max": 0.007195723708719015,
      "clip_ratio/high_mean": 0.0035978618543595076,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035978618543595076,
      "entropy": 0.5177513211965561,
      "epoch": 8e-05,
      "grad_norm": 2.218416929244995,
      "kl": 0.0013052129361312836,
      "learning_rate": 8.571428571428572e-07,
      "loss": 0.1144,
      "step": 4,
      "step_time": 6.260843116993783
    },
    {
      "clip_ratio/high_max": 0.0036764706019312143,
      "clip_ratio/high_mean": 0.0018382353009656072,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0018382353009656072,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1892.0,
      "completions/max_terminated_length": 1892.0,
      "completions/mean_length": 1607.125,
      "completions/mean_terminated_length": 1607.125,
      "completions/min_length": 274.0,
      "completions/min_terminated_length": 274.0,
      "entropy": 0.5215674266219139,
      "epoch": 0.0001,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 2.6341018676757812,
      "kl": 0.0011459336819825694,
      "learning_rate": 1.142857142857143e-06,
      "loss": -0.0192,
      "num_tokens": 226943.0,
      "reward": -0.6061195731163025,
      "reward_std": 0.2195071578025818,
      "rewards/rollout_reward_func/mean": -0.6061195731163025,
      "rewards/rollout_reward_func/std": 0.23115375638008118,
      "sampling/importance_sampling_ratio/max": 2.042933940887451,
      "sampling/importance_sampling_ratio/mean": 0.9376565217971802,
      "sampling/importance_sampling_ratio/min": 3.4871042007720773e-14,
      "sampling/sampling_logp_difference/max": 28.74315643310547,
      "sampling/sampling_logp_difference/mean": 0.14062589406967163,
      "step": 5,
      "step_time": 31.375892003994522
    },
    {
      "clip_ratio/high_max": 0.007352941203862429,
      "clip_ratio/high_mean": 0.0036764706019312143,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0036764706019312143,
      "entropy": 0.5232524983584881,
      "epoch": 0.00012,
      "grad_norm": 2.3644168376922607,
      "kl": 0.0010918559346464463,
      "learning_rate": 1.4285714285714286e-06,
      "loss": -0.0195,
      "step": 6,
      "step_time": 6.260178877000726
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0016025641234591603,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0035556891234591603,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1890.0,
      "completions/max_terminated_length": 1890.0,
      "completions/mean_length": 1565.90625,
      "completions/mean_terminated_length": 1565.90625,
      "completions/min_length": 492.0,
      "completions/min_terminated_length": 492.0,
      "entropy": 0.48773781582713127,
      "epoch": 0.00014,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.9083935022354126,
      "kl": 0.0010215075744781643,
      "learning_rate": 1.7142857142857145e-06,
      "loss": 0.0341,
      "num_tokens": 300553.0,
      "reward": -0.6151305437088013,
      "reward_std": 0.23046068847179413,
      "rewards/rollout_reward_func/mean": -0.6151305437088013,
      "rewards/rollout_reward_func/std": 0.24246515333652496,
      "sampling/importance_sampling_ratio/max": 1.3417084217071533,
      "sampling/importance_sampling_ratio/mean": 0.945120632648468,
      "sampling/importance_sampling_ratio/min": 2.504423404769973e-11,
      "sampling/sampling_logp_difference/max": 24.4301700592041,
      "sampling/sampling_logp_difference/mean": 0.06543836742639542,
      "step": 7,
      "step_time": 30.44348550399809
    },
    {
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.003388278419151902,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.003388278419151902,
      "entropy": 0.48545313626527786,
      "epoch": 0.00016,
      "grad_norm": 1.8554192781448364,
      "kl": 0.0009212387376464903,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0347,
      "step": 8,
      "step_time": 6.283500026991533
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.00533150346018374,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00728462846018374,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1925.0,
      "completions/max_terminated_length": 1925.0,
      "completions/mean_length": 1673.8125,
      "completions/mean_terminated_length": 1673.8125,
      "completions/min_length": 1305.0,
      "completions/min_terminated_length": 1305.0,
      "entropy": 0.5548458024859428,
      "epoch": 0.00018,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.4972906112670898,
      "kl": 0.001124227695981972,
      "learning_rate": 2.285714285714286e-06,
      "loss": -0.0231,
      "num_tokens": 377565.0,
      "reward": -0.606047511100769,
      "reward_std": 0.17791865766048431,
      "rewards/rollout_reward_func/mean": -0.606047511100769,
      "rewards/rollout_reward_func/std": 0.21615628898143768,
      "sampling/importance_sampling_ratio/max": 1.6678369045257568,
      "sampling/importance_sampling_ratio/mean": 0.9152076244354248,
      "sampling/importance_sampling_ratio/min": 4.311538348567012e-10,
      "sampling/sampling_logp_difference/max": 20.632226943969727,
      "sampling/sampling_logp_difference/mean": 0.06397730857133865,
      "step": 9,
      "step_time": 41.25972091700169
    },
    {
      "clip_ratio/high_max": 0.00390625,
      "clip_ratio/high_mean": 0.001953125,
      "clip_ratio/low_mean": 0.0016891892300918698,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.00364231423009187,
      "entropy": 0.5546580441296101,
      "epoch": 0.0002,
      "grad_norm": 1.610571265220642,
      "kl": 0.001391250429151114,
      "learning_rate": 2.571428571428571e-06,
      "loss": -0.0255,
      "step": 10,
      "step_time": 7.209359069009224
    },
    {
      "clip_ratio/high_max": 0.009114583488553762,
      "clip_ratio/high_mean": 0.004557291744276881,
      "clip_ratio/low_mean": 0.0022321429569274187,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.006789434468373656,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1808.0,
      "completions/max_terminated_length": 1808.0,
      "completions/mean_length": 1554.34375,
      "completions/mean_terminated_length": 1554.34375,
      "completions/min_length": 700.0,
      "completions/min_terminated_length": 700.0,
      "entropy": 0.4812197834253311,
      "epoch": 0.00022,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 1.6533305644989014,
      "kl": 0.0016009878600016236,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 0.0432,
      "num_tokens": 451126.0,
      "reward": -0.5563069581985474,
      "reward_std": 0.29338371753692627,
      "rewards/rollout_reward_func/mean": -0.5563069581985474,
      "rewards/rollout_reward_func/std": 0.3663959503173828,
      "sampling/importance_sampling_ratio/max": 1.6875137090682983,
      "sampling/importance_sampling_ratio/mean": 1.0345326662063599,
      "sampling/importance_sampling_ratio/min": 1.5839532376688004e-13,
      "sampling/sampling_logp_difference/max": 28.526655197143555,
      "sampling/sampling_logp_difference/mean": 0.07865896075963974,
      "step": 11,
      "step_time": 34.2954730700003
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 350000,
  "num_input_tokens_seen": 451126,
  "num_train_epochs": 7,
  "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
}