trainer_state.json

{
  "best_metric": 0.9112,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0559/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": 1.2082713842391968,
      "learning_rate": 0.0004849231551964771,
      "loss": 0.778,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8192,
      "eval_loss": 0.6414018273353577,
      "eval_runtime": 10.5168,
      "eval_samples_per_second": 356.574,
      "eval_steps_per_second": 5.61,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.3052964210510254,
      "learning_rate": 0.0004415111107797445,
      "loss": 0.4327,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8314666666666667,
      "eval_loss": 0.5531339645385742,
      "eval_runtime": 10.3693,
      "eval_samples_per_second": 361.644,
      "eval_steps_per_second": 5.69,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.07922560721635818,
      "learning_rate": 0.000375,
      "loss": 0.2874,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8490666666666666,
      "eval_loss": 0.5388678908348083,
      "eval_runtime": 10.2794,
      "eval_samples_per_second": 364.809,
      "eval_steps_per_second": 5.74,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0002936701006654613,
      "loss": 0.1899,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8522666666666666,
      "eval_loss": 0.5192283987998962,
      "eval_runtime": 10.3576,
      "eval_samples_per_second": 362.052,
      "eval_steps_per_second": 5.696,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.15561629831790924,
      "learning_rate": 0.00020684605953398183,
      "loss": 0.092,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8658666666666667,
      "eval_loss": 0.5275271534919739,
      "eval_runtime": 10.7389,
      "eval_samples_per_second": 349.199,
      "eval_steps_per_second": 5.494,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.015237436629831791,
      "learning_rate": 0.00012522702050703539,
      "loss": 0.0313,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8954666666666666,
      "eval_loss": 0.44514814019203186,
      "eval_runtime": 11.484,
      "eval_samples_per_second": 326.541,
      "eval_steps_per_second": 5.138,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.04188411682844162,
      "learning_rate": 5.8657444259997555e-05,
      "loss": 0.0089,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9082666666666667,
      "eval_loss": 0.39722031354904175,
      "eval_runtime": 10.6434,
      "eval_samples_per_second": 352.33,
      "eval_steps_per_second": 5.543,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.10371667146682739,
      "learning_rate": 1.5166604150571307e-05,
      "loss": 0.0022,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9072,
      "eval_loss": 0.39766111969947815,
      "eval_runtime": 10.4483,
      "eval_samples_per_second": 358.911,
      "eval_steps_per_second": 5.647,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.004805553704500198,
      "learning_rate": 1.373523937919785e-10,
      "loss": 0.0014,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9112,
      "eval_loss": 0.39600107073783875,
      "eval_runtime": 10.5425,
      "eval_samples_per_second": 355.704,
      "eval_steps_per_second": 5.596,
      "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
}