trainer_state.json

{
  "best_metric": 0.7714666666666666,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0000/checkpoints/checkpoint-999",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 999,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 225.25152587890625,
      "learning_rate": 3e-05,
      "loss": 3.8339,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4298666666666667,
      "eval_loss": 3.687291383743286,
      "eval_runtime": 8.7325,
      "eval_samples_per_second": 429.431,
      "eval_steps_per_second": 6.756,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 440.04058837890625,
      "learning_rate": 3e-05,
      "loss": 2.9454,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.636,
      "eval_loss": 1.8807346820831299,
      "eval_runtime": 8.6784,
      "eval_samples_per_second": 432.109,
      "eval_steps_per_second": 6.799,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 64.37602996826172,
      "learning_rate": 3e-05,
      "loss": 1.3544,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7714666666666666,
      "eval_loss": 0.9381110668182373,
      "eval_runtime": 8.5885,
      "eval_samples_per_second": 436.629,
      "eval_steps_per_second": 6.87,
      "step": 999
    }
  ],
  "logging_steps": 500,
  "max_steps": 999,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.4529287718656e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}