checkpoint-680/trainer_state.json

{
  "best_metric": 0.8267140984535217,
  "best_model_checkpoint": "face_emotions_image_detection/checkpoint-680",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 680,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7641704592470004,
      "eval_loss": 0.8311234712600708,
      "eval_runtime": 36.6909,
      "eval_samples_per_second": 65.875,
      "eval_steps_per_second": 2.071,
      "step": 340
    },
    {
      "epoch": 1.47,
      "grad_norm": 2.571842670440674,
      "learning_rate": 1.4285714285714285e-07,
      "loss": 0.7492,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7641704592470004,
      "eval_loss": 0.8267140984535217,
      "eval_runtime": 36.0762,
      "eval_samples_per_second": 66.997,
      "eval_steps_per_second": 2.107,
      "step": 680
    }
  ],
  "logging_steps": 500,
  "max_steps": 680,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 3.3702474498545664e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}