trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.764705882352941,
  "eval_steps": 25,
  "global_step": 8,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.47058823529411764,
      "grad_norm": 1.3573589324951172,
      "learning_rate": 0.000175,
      "loss": 9.573,
      "step": 1
    },
    {
      "epoch": 0.9411764705882353,
      "grad_norm": 1.2206748723983765,
      "learning_rate": 0.00015000000000000001,
      "loss": 9.4647,
      "step": 2
    },
    {
      "epoch": 1.4117647058823528,
      "grad_norm": 1.1659976243972778,
      "learning_rate": 0.000125,
      "loss": 9.2541,
      "step": 3
    },
    {
      "epoch": 1.8823529411764706,
      "grad_norm": 1.3362257480621338,
      "learning_rate": 0.0001,
      "loss": 9.0185,
      "step": 4
    },
    {
      "epoch": 2.3529411764705883,
      "grad_norm": 1.568061351776123,
      "learning_rate": 7.500000000000001e-05,
      "loss": 8.8428,
      "step": 5
    },
    {
      "epoch": 2.8235294117647056,
      "grad_norm": 1.6825159788131714,
      "learning_rate": 5e-05,
      "loss": 8.7436,
      "step": 6
    },
    {
      "epoch": 3.2941176470588234,
      "grad_norm": 2.018841028213501,
      "learning_rate": 2.5e-05,
      "loss": 8.5273,
      "step": 7
    },
    {
      "epoch": 3.764705882352941,
      "grad_norm": 2.155460834503174,
      "learning_rate": 0.0,
      "loss": 8.483,
      "step": 8
    },
    {
      "epoch": 3.764705882352941,
      "step": 8,
      "total_flos": 35328595626816.0,
      "train_loss": 8.98837435245514,
      "train_runtime": 239.7476,
      "train_samples_per_second": 0.551,
      "train_steps_per_second": 0.033
    }
  ],
  "logging_steps": 1,
  "max_steps": 8,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 35328595626816.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}