trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 65,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "learning_rate": 0.00016923076923076923,
      "loss": 0.184,
      "step": 10
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00014153846153846156,
      "loss": 0.2527,
      "step": 20
    },
    {
      "epoch": 0.46,
      "learning_rate": 0.00011076923076923077,
      "loss": 0.3185,
      "step": 30
    },
    {
      "epoch": 0.62,
      "learning_rate": 8e-05,
      "loss": 0.176,
      "step": 40
    },
    {
      "epoch": 0.77,
      "learning_rate": 4.923076923076924e-05,
      "loss": 0.1369,
      "step": 50
    },
    {
      "epoch": 0.92,
      "learning_rate": 1.8461538461538465e-05,
      "loss": 0.1163,
      "step": 60
    },
    {
      "epoch": 1.0,
      "step": 65,
      "total_flos": 8.012743541191066e+16,
      "train_loss": 0.1892859371808859,
      "train_runtime": 26.1148,
      "train_samples_per_second": 39.594,
      "train_steps_per_second": 2.489
    }
  ],
  "max_steps": 65,
  "num_train_epochs": 1,
  "total_flos": 8.012743541191066e+16,
  "trial_name": null,
  "trial_params": null
}