{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00011,
  "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": 0.0625,
      "completions/max_length": 1752.0,
      "completions/max_terminated_length": 1752.0,
      "completions/mean_length": 1533.40625,
      "completions/mean_terminated_length": 1526.3333740234375,
      "completions/min_length": 1257.0,
      "completions/min_terminated_length": 1257.0,
      "entropy": 0.3670750502496958,
      "epoch": 1e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 5.134232521057129,
      "kl": 0.0,
      "learning_rate": 0.0,
      "loss": 0.5565,
      "num_tokens": 70741.0,
      "reward": -2.7539215087890625,
      "reward_std": 8.533812522888184,
      "rewards/rollout_reward_func/mean": -2.7539215087890625,
      "rewards/rollout_reward_func/std": 9.058825492858887,
      "sampling/importance_sampling_ratio/max": 2.21063232421875,
      "sampling/importance_sampling_ratio/mean": 0.9861273765563965,
      "sampling/importance_sampling_ratio/min": 0.0,
      "sampling/sampling_logp_difference/max": 15.751568794250488,
      "sampling/sampling_logp_difference/mean": 0.06819069385528564,
      "step": 1,
      "step_time": 37.74923437399957
    },
    {
      "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": 0.3670750502496958,
      "epoch": 2e-05,
      "grad_norm": 5.1552276611328125,
      "kl": 0.0,
      "learning_rate": 2.8571428571428575e-07,
      "loss": 0.5565,
      "step": 2,
      "step_time": 6.120819243000369
    },
    {
      "clip_ratio/high_max": 0.017710780957713723,
      "clip_ratio/high_mean": 0.009723446099087596,
      "clip_ratio/low_mean": 0.0025572447921149433,
      "clip_ratio/low_min": 0.0016891892300918698,
      "clip_ratio/region_mean": 0.012280690658371896,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1780.0,
      "completions/max_terminated_length": 1780.0,
      "completions/mean_length": 1557.40625,
      "completions/mean_terminated_length": 1557.40625,
      "completions/min_length": 1426.0,
      "completions/min_terminated_length": 1426.0,
      "entropy": 0.38403771445155144,
      "epoch": 3e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 4.518792152404785,
      "kl": 0.003740338608622551,
      "learning_rate": 5.714285714285715e-07,
      "loss": 0.1169,
      "num_tokens": 142373.0,
      "reward": -0.3596813678741455,
      "reward_std": 4.564228534698486,
      "rewards/rollout_reward_func/mean": -0.3596813678741455,
      "rewards/rollout_reward_func/std": 10.295997619628906,
      "sampling/importance_sampling_ratio/max": 2.081014394760132,
      "sampling/importance_sampling_ratio/mean": 0.9942376613616943,
      "sampling/importance_sampling_ratio/min": 5.060187137039498e-36,
      "sampling/sampling_logp_difference/max": 29.89274787902832,
      "sampling/sampling_logp_difference/mean": 0.13873185217380524,
      "step": 3,
      "step_time": 36.35921934200087
    },
    {
      "clip_ratio/high_max": 0.011207867413759232,
      "clip_ratio/high_mean": 0.00647198932711035,
      "clip_ratio/low_mean": 0.00706147204618901,
      "clip_ratio/low_min": 0.0017361111240461469,
      "clip_ratio/region_mean": 0.013533461140468717,
      "entropy": 0.38358503952622414,
      "epoch": 4e-05,
      "grad_norm": 4.347132205963135,
      "kl": 0.011755978513974696,
      "learning_rate": 8.571428571428572e-07,
      "loss": 0.119,
      "step": 4,
      "step_time": 6.707314684997982
    },
    {
      "clip_ratio/high_max": 0.012880883645266294,
      "clip_ratio/high_mean": 0.008077950216829777,
      "clip_ratio/low_mean": 0.005651012295857072,
      "clip_ratio/low_min": 0.0016891892300918698,
      "clip_ratio/region_mean": 0.013728962745517492,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1757.0,
      "completions/max_terminated_length": 1757.0,
      "completions/mean_length": 1573.75,
      "completions/mean_terminated_length": 1573.75,
      "completions/min_length": 1443.0,
      "completions/min_terminated_length": 1443.0,
      "entropy": 0.3456532806158066,
      "epoch": 5e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 4.78507661819458,
      "kl": 0.004491248982958496,
      "learning_rate": 1.142857142857143e-06,
      "loss": -0.0491,
      "num_tokens": 214405.0,
      "reward": -2.5846972465515137,
      "reward_std": 9.13107681274414,
      "rewards/rollout_reward_func/mean": -2.5846972465515137,
      "rewards/rollout_reward_func/std": 12.559075355529785,
      "sampling/importance_sampling_ratio/max": 2.6732370853424072,
      "sampling/importance_sampling_ratio/mean": 0.8290778398513794,
      "sampling/importance_sampling_ratio/min": 4.706499113173833e-40,
      "sampling/sampling_logp_difference/max": 28.09636878967285,
      "sampling/sampling_logp_difference/mean": 0.18767967820167542,
      "step": 5,
      "step_time": 34.56705875800071
    },
    {
      "clip_ratio/high_max": 0.015001210267655551,
      "clip_ratio/high_mean": 0.007500605133827776,
      "clip_ratio/low_mean": 0.010737838223576546,
      "clip_ratio/low_min": 0.004310344811528921,
      "clip_ratio/region_mean": 0.018238443473819643,
      "entropy": 0.345603134483099,
      "epoch": 6e-05,
      "grad_norm": 4.702767372131348,
      "kl": 0.00609777684439905,
      "learning_rate": 1.4285714285714286e-06,
      "loss": -0.0462,
      "step": 6,
      "step_time": 6.165537762000895
    },
    {
      "clip_ratio/high_max": 0.026439692708663642,
      "clip_ratio/high_mean": 0.01679553568828851,
      "clip_ratio/low_mean": 0.00500575453042984,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.021801290102303028,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 1775.0,
      "completions/max_terminated_length": 1701.0,
      "completions/mean_length": 1541.125,
      "completions/mean_terminated_length": 1525.7333984375,
      "completions/min_length": 1423.0,
      "completions/min_terminated_length": 1423.0,
      "entropy": 0.39175233244895935,
      "epoch": 7e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 12.207725524902344,
      "kl": 0.007055700887576677,
      "learning_rate": 1.7142857142857145e-06,
      "loss": 0.1615,
      "num_tokens": 284511.0,
      "reward": -8.317660331726074,
      "reward_std": 4.3451619148254395,
      "rewards/rollout_reward_func/mean": -8.317660331726074,
      "rewards/rollout_reward_func/std": 4.4500837326049805,
      "sampling/importance_sampling_ratio/max": 1.7571941614151,
      "sampling/importance_sampling_ratio/mean": 0.8103891611099243,
      "sampling/importance_sampling_ratio/min": 1.6684203226478317e-26,
      "sampling/sampling_logp_difference/max": 24.94557762145996,
      "sampling/sampling_logp_difference/mean": 0.07305071502923965,
      "step": 7,
      "step_time": 35.98801361099686
    },
    {
      "clip_ratio/high_max": 0.02302091906312853,
      "clip_ratio/high_mean": 0.012378515035379678,
      "clip_ratio/low_mean": 0.006711377413012087,
      "clip_ratio/low_min": 0.0031250000465661287,
      "clip_ratio/region_mean": 0.019089892564807087,
      "entropy": 0.39244432002305984,
      "epoch": 8e-05,
      "grad_norm": 5.381319522857666,
      "kl": 0.01739203451143112,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.1543,
      "step": 8,
      "step_time": 6.732931664002535
    },
    {
      "clip_ratio/high_max": 0.012895698891952634,
      "clip_ratio/high_mean": 0.008981633465737104,
      "clip_ratio/low_mean": 0.008420342986937612,
      "clip_ratio/low_min": 0.0014534883666783571,
      "clip_ratio/region_mean": 0.01740197604522109,
      "completions/clipped_ratio": 0.0625,
      "completions/max_length": 1966.0,
      "completions/max_terminated_length": 1855.0,
      "completions/mean_length": 1668.375,
      "completions/mean_terminated_length": 1649.5667724609375,
      "completions/min_length": 1461.0,
      "completions/min_terminated_length": 1461.0,
      "entropy": 0.3468770608305931,
      "epoch": 9e-05,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 2.7011311054229736,
      "kl": 0.004557034379104152,
      "learning_rate": 2.285714285714286e-06,
      "loss": 0.2606,
      "num_tokens": 359632.0,
      "reward": -4.431756019592285,
      "reward_std": 7.569368362426758,
      "rewards/rollout_reward_func/mean": -4.431756019592285,
      "rewards/rollout_reward_func/std": 7.730715274810791,
      "sampling/importance_sampling_ratio/max": 2.27473521232605,
      "sampling/importance_sampling_ratio/mean": 1.0028448104858398,
      "sampling/importance_sampling_ratio/min": 1.6143619129999664e-21,
      "sampling/sampling_logp_difference/max": 28.98799705505371,
      "sampling/sampling_logp_difference/mean": 0.07533948123455048,
      "step": 9,
      "step_time": 38.395322346001194
    },
    {
      "clip_ratio/high_max": 0.0125588474329561,
      "clip_ratio/high_mean": 0.009703430347144604,
      "clip_ratio/low_mean": 0.0093731161323376,
      "clip_ratio/low_min": 0.0016891892300918698,
      "clip_ratio/region_mean": 0.019076546072028577,
      "entropy": 0.34513476490974426,
      "epoch": 0.0001,
      "grad_norm": 2.967703104019165,
      "kl": 0.00513960600073915,
      "learning_rate": 2.571428571428571e-06,
      "loss": 0.2603,
      "step": 10,
      "step_time": 6.693931524998334
    },
    {
      "clip_ratio/high_max": 0.02302549034357071,
      "clip_ratio/high_mean": 0.01321350410580635,
      "clip_ratio/low_mean": 0.010665920155588537,
      "clip_ratio/low_min": 0.0035211266949772835,
      "clip_ratio/region_mean": 0.023879424086771905,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1776.0,
      "completions/max_terminated_length": 1776.0,
      "completions/mean_length": 1566.75,
      "completions/mean_terminated_length": 1566.75,
      "completions/min_length": 1419.0,
      "completions/min_terminated_length": 1419.0,
      "entropy": 0.40919550508260727,
      "epoch": 0.00011,
      "frac_reward_zero_std": 0.0,
      "grad_norm": 4.4168381690979,
      "kl": 0.008918058272683993,
      "learning_rate": 2.8571428571428573e-06,
      "loss": -0.3278,
      "num_tokens": 430181.0,
      "reward": -0.3421344757080078,
      "reward_std": 7.553745269775391,
      "rewards/rollout_reward_func/mean": -0.3421344757080078,
      "rewards/rollout_reward_func/std": 11.30397891998291,
      "sampling/importance_sampling_ratio/max": 2.6663432121276855,
      "sampling/importance_sampling_ratio/mean": 0.8930033445358276,
      "sampling/importance_sampling_ratio/min": 0.0,
      "sampling/sampling_logp_difference/max": 29.21600341796875,
      "sampling/sampling_logp_difference/mean": 0.22408267855644226,
      "step": 11,
      "step_time": 35.09687606099942
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 500000,
  "num_input_tokens_seen": 430181,
  "num_train_epochs": 5,
  "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
}