{
  "best_global_step": 10,
  "best_metric": 1.440527398699487e-06,
  "best_model_checkpoint": null,
  "epoch": 2.96,
  "eval_steps": 5,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8,
      "grad_norm": 0.025135431438684464,
      "learning_rate": 7.777777777777778e-07,
      "loss": 0.0025,
      "step": 5
    },
    {
      "epoch": 0.8,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 250.1875,
      "eval_kl": 4.038810729980469e-05,
      "eval_loss": 0.01284346729516983,
      "eval_num_tokens": 31676.0,
      "eval_reward": 0.025,
      "eval_reward_std": 0.05,
      "eval_rewards/equation_reward_func": 0.0125,
      "eval_rewards/format_reward_func": 0.0125,
      "eval_runtime": 283.8664,
      "eval_samples_per_second": 0.07,
      "eval_steps_per_second": 0.018,
      "step": 5
    },
    {
      "epoch": 1.6400000000000001,
      "grad_norm": 0.0,
      "learning_rate": 5e-07,
      "loss": -0.0027,
      "step": 10
    },
    {
      "epoch": 1.6400000000000001,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 251.9,
      "eval_kl": 3.6042928695678714e-05,
      "eval_loss": 1.440527398699487e-06,
      "eval_num_tokens": 63054.0,
      "eval_reward": 0.0125,
      "eval_reward_std": 0.025,
      "eval_rewards/equation_reward_func": 0.0125,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 283.8833,
      "eval_samples_per_second": 0.07,
      "eval_steps_per_second": 0.018,
      "step": 10
    },
    {
      "epoch": 2.48,
      "grad_norm": 1.2480842087825295e-05,
      "learning_rate": 2.222222222222222e-07,
      "loss": 0.0025,
      "step": 15
    },
    {
      "epoch": 2.48,
      "eval_clip_ratio": 0.0,
      "eval_completion_length": 252.1875,
      "eval_kl": 4.0525197982788085e-05,
      "eval_loss": 0.004817621782422066,
      "eval_num_tokens": 94681.0,
      "eval_reward": 0.05,
      "eval_reward_std": 0.1,
      "eval_rewards/equation_reward_func": 0.05,
      "eval_rewards/format_reward_func": 0.0,
      "eval_runtime": 283.0601,
      "eval_samples_per_second": 0.071,
      "eval_steps_per_second": 0.018,
      "step": 15
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 15,
  "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": 4,
  "trial_name": null,
  "trial_params": null
}