trainer_state.json

{
  "best_metric": 0.8933333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0294/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": Infinity,
      "learning_rate": 8.003003003003004e-05,
      "loss": 1.5757,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8013333333333333,
      "eval_loss": 0.7107900977134705,
      "eval_runtime": 6.8737,
      "eval_samples_per_second": 545.562,
      "eval_steps_per_second": 8.584,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.44568052887916565,
      "learning_rate": 7.003003003003004e-05,
      "loss": 0.4274,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8573333333333333,
      "eval_loss": 0.4989762306213379,
      "eval_runtime": 6.1583,
      "eval_samples_per_second": 608.936,
      "eval_steps_per_second": 9.581,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 14.441655158996582,
      "learning_rate": 6.003003003003003e-05,
      "loss": 0.2059,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8578666666666667,
      "eval_loss": 0.49824610352516174,
      "eval_runtime": 6.2591,
      "eval_samples_per_second": 599.124,
      "eval_steps_per_second": 9.426,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.23665133118629456,
      "learning_rate": 5.003003003003004e-05,
      "loss": 0.1018,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8608,
      "eval_loss": 0.5123676061630249,
      "eval_runtime": 6.3285,
      "eval_samples_per_second": 592.553,
      "eval_steps_per_second": 9.323,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.048037879168987274,
      "learning_rate": 4.003003003003003e-05,
      "loss": 0.0497,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8658666666666667,
      "eval_loss": 0.5245042443275452,
      "eval_runtime": 6.4054,
      "eval_samples_per_second": 585.443,
      "eval_steps_per_second": 9.211,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.0858592838048935,
      "learning_rate": 3.003003003003003e-05,
      "loss": 0.022,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8856,
      "eval_loss": 0.4839324355125427,
      "eval_runtime": 6.2762,
      "eval_samples_per_second": 597.491,
      "eval_steps_per_second": 9.401,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.030141333118081093,
      "learning_rate": 2.0030030030030034e-05,
      "loss": 0.0077,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8866666666666667,
      "eval_loss": 0.49923765659332275,
      "eval_runtime": 6.0304,
      "eval_samples_per_second": 621.845,
      "eval_steps_per_second": 9.784,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.11882772296667099,
      "learning_rate": 1.0030030030030031e-05,
      "loss": 0.0031,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8909333333333334,
      "eval_loss": 0.4866844117641449,
      "eval_runtime": 6.1255,
      "eval_samples_per_second": 612.19,
      "eval_steps_per_second": 9.632,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.037997767329216,
      "learning_rate": 3.003003003003003e-08,
      "loss": 0.0018,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8933333333333333,
      "eval_loss": 0.4823809266090393,
      "eval_runtime": 6.2072,
      "eval_samples_per_second": 604.132,
      "eval_steps_per_second": 9.505,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}