{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 1,
  "global_step": 10,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.472108991167676,
      "learning_rate": 2e-05,
      "loss": 0.0368,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 573.25,
      "eval_kl": 0.0,
      "eval_loss": 0.022564562037587166,
      "eval_reward": 0.25,
      "eval_reward_std": 0.28867512941360474,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.125,
      "eval_rewards/format_reward_staging": 0.125,
      "eval_runtime": 10.3415,
      "eval_samples_per_second": 1.547,
      "eval_steps_per_second": 0.097,
      "step": 1
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5030364342093353,
      "learning_rate": 1.9396926207859085e-05,
      "loss": -0.0324,
      "step": 2
    },
    {
      "epoch": 2.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 476.625,
      "eval_kl": 2.46875,
      "eval_loss": 0.14496715366840363,
      "eval_reward": 0.75,
      "eval_reward_std": 0.7673887014389038,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.375,
      "eval_rewards/format_reward_staging": 0.375,
      "eval_runtime": 9.814,
      "eval_samples_per_second": 1.63,
      "eval_steps_per_second": 0.102,
      "step": 2
    },
    {
      "epoch": 3.0,
      "grad_norm": 23.95943994129872,
      "learning_rate": 1.766044443118978e-05,
      "loss": 0.0685,
      "step": 3
    },
    {
      "epoch": 3.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 544.375,
      "eval_kl": 63.25,
      "eval_loss": 1.4514538049697876,
      "eval_reward": 1.125,
      "eval_reward_std": 0.7886751294136047,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.625,
      "eval_rewards/format_reward_staging": 0.5,
      "eval_runtime": 10.8975,
      "eval_samples_per_second": 1.468,
      "eval_steps_per_second": 0.092,
      "step": 3
    },
    {
      "epoch": 4.0,
      "grad_norm": 40.3897041999098,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.1064,
      "step": 4
    },
    {
      "epoch": 4.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 490.1875,
      "eval_kl": 1.453125,
      "eval_loss": 0.18356087803840637,
      "eval_reward": 1.375,
      "eval_reward_std": 0.7427992820739746,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.75,
      "eval_rewards/format_reward_staging": 0.625,
      "eval_runtime": 9.682,
      "eval_samples_per_second": 1.653,
      "eval_steps_per_second": 0.103,
      "step": 4
    },
    {
      "epoch": 5.0,
      "grad_norm": 617.8779151304285,
      "learning_rate": 1.1736481776669307e-05,
      "loss": 0.4202,
      "step": 5
    },
    {
      "epoch": 5.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 579.5,
      "eval_kl": 0.208984375,
      "eval_loss": 0.23663775622844696,
      "eval_reward": 1.25,
      "eval_reward_std": 0.9082483053207397,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.75,
      "eval_rewards/format_reward_staging": 0.5,
      "eval_runtime": 10.2564,
      "eval_samples_per_second": 1.56,
      "eval_steps_per_second": 0.097,
      "step": 5
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9504042286680588,
      "learning_rate": 8.263518223330698e-06,
      "loss": 0.0833,
      "step": 6
    },
    {
      "epoch": 6.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 532.4375,
      "eval_kl": 0.208984375,
      "eval_loss": 0.25171786546707153,
      "eval_reward": 1.0,
      "eval_reward_std": 0.8643567562103271,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.5,
      "eval_rewards/format_reward_staging": 0.5,
      "eval_runtime": 10.0367,
      "eval_samples_per_second": 1.594,
      "eval_steps_per_second": 0.1,
      "step": 6
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.8737839403793075,
      "learning_rate": 5.000000000000003e-06,
      "loss": 0.1363,
      "step": 7
    },
    {
      "epoch": 7.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 618.0,
      "eval_kl": 0.12353515625,
      "eval_loss": 0.12567630410194397,
      "eval_reward": 0.875,
      "eval_reward_std": 0.5684809684753418,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.625,
      "eval_rewards/format_reward_staging": 0.25,
      "eval_runtime": 11.9883,
      "eval_samples_per_second": 1.335,
      "eval_steps_per_second": 0.083,
      "step": 7
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.9960909069556698,
      "learning_rate": 2.339555568810221e-06,
      "loss": 0.1077,
      "step": 8
    },
    {
      "epoch": 8.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 522.5625,
      "eval_kl": 0.220703125,
      "eval_loss": 0.17136843502521515,
      "eval_reward": 1.125,
      "eval_reward_std": 0.7427992820739746,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.75,
      "eval_rewards/format_reward_staging": 0.375,
      "eval_runtime": 9.9314,
      "eval_samples_per_second": 1.611,
      "eval_steps_per_second": 0.101,
      "step": 8
    },
    {
      "epoch": 9.0,
      "grad_norm": 1.4868290469557222,
      "learning_rate": 6.030737921409169e-07,
      "loss": 0.0948,
      "step": 9
    },
    {
      "epoch": 9.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 449.9375,
      "eval_kl": 0.2080078125,
      "eval_loss": 0.18853142857551575,
      "eval_reward": 1.625,
      "eval_reward_std": 0.4330126941204071,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.875,
      "eval_rewards/format_reward_staging": 0.75,
      "eval_runtime": 9.7136,
      "eval_samples_per_second": 1.647,
      "eval_steps_per_second": 0.103,
      "step": 9
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.247559086562991,
      "learning_rate": 0.0,
      "loss": 0.1509,
      "step": 10
    },
    {
      "epoch": 10.0,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 586.6875,
      "eval_kl": 0.1611328125,
      "eval_loss": 0.21008853614330292,
      "eval_reward": 1.0,
      "eval_reward_std": 0.843070387840271,
      "eval_rewards/accuracy_reward_math": 0.0,
      "eval_rewards/accuracy_reward_staging": 0.0,
      "eval_rewards/format_reward": 0.6875,
      "eval_rewards/format_reward_staging": 0.3125,
      "eval_runtime": 9.7606,
      "eval_samples_per_second": 1.639,
      "eval_steps_per_second": 0.102,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 563.915625,
      "epoch": 10.0,
      "kl": 0.9728057861328125,
      "reward": 3.8234375,
      "reward_std": 4.318631359934807,
      "rewards/accuracy_reward_math": 0.028125,
      "rewards/accuracy_reward_staging": 0.0,
      "rewards/format_reward": 0.6015625,
      "rewards/format_reward_staging": 0.409375,
      "step": 10,
      "total_flos": 0.0,
      "train_loss": 0.11724991016089917,
      "train_runtime": 482.6536,
      "train_samples_per_second": 0.332,
      "train_steps_per_second": 0.021
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 200,
  "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": 16,
  "trial_name": null,
  "trial_params": null
}