trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.1,
  "eval_steps": 282,
  "global_step": 564,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0007092198581560284,
      "grad_norm": 26.586772918701172,
      "learning_rate": 1e-05,
      "loss": 4.8407,
      "step": 1
    },
    {
      "epoch": 1.1,
      "grad_norm": 0.9762190580368042,
      "learning_rate": 0.0008610687022900763,
      "loss": 1.5428,
      "step": 282
    },
    {
      "epoch": 1.1,
      "eval_loss": 1.2981462478637695,
      "eval_runtime": 49.4189,
      "eval_samples_per_second": 9.086,
      "eval_steps_per_second": 1.153,
      "step": 282
    },
    {
      "epoch": 3.1,
      "grad_norm": 1.0431804656982422,
      "learning_rate": 0.0006458015267175574,
      "loss": 0.7498,
      "step": 564
    },
    {
      "epoch": 3.1,
      "eval_loss": 0.8457677960395813,
      "eval_runtime": 49.1336,
      "eval_samples_per_second": 9.138,
      "eval_steps_per_second": 1.16,
      "step": 564
    }
  ],
  "logging_steps": 282,
  "max_steps": 1410,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 282,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.37203291897856e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}