trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 7.9406179427600705,
      "learning_rate": 2.5e-06,
      "loss": 1.4592,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 4.421267130268739,
      "learning_rate": 4.3712768704277535e-06,
      "loss": 1.3802,
      "step": 5
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.3249047994613647,
      "eval_runtime": 4.0558,
      "eval_samples_per_second": 11.095,
      "eval_steps_per_second": 0.74,
      "step": 5
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.5327777442940533,
      "learning_rate": 1.613487782393661e-06,
      "loss": 1.1617,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.2391588687896729,
      "eval_runtime": 4.1583,
      "eval_samples_per_second": 10.822,
      "eval_steps_per_second": 0.721,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.902024722630514,
      "learning_rate": 0.0,
      "loss": 1.0651,
      "step": 15
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.208680510520935,
      "eval_runtime": 4.1534,
      "eval_samples_per_second": 10.834,
      "eval_steps_per_second": 0.722,
      "step": 15
    },
    {
      "epoch": 3.0,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.2075830618540446,
      "train_runtime": 257.198,
      "train_samples_per_second": 3.663,
      "train_steps_per_second": 0.058
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3088349921280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}