trainer_state.json

{
  "best_metric": 0.9496,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0758/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": 5.763750076293945,
      "learning_rate": 6.222222222222222e-05,
      "loss": 0.6973,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9181333333333334,
      "eval_loss": 0.2694110870361328,
      "eval_runtime": 11.2616,
      "eval_samples_per_second": 332.991,
      "eval_steps_per_second": 5.239,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.1675330400466919,
      "learning_rate": 5.444444444444444e-05,
      "loss": 0.1426,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9314666666666667,
      "eval_loss": 0.22955980896949768,
      "eval_runtime": 11.4223,
      "eval_samples_per_second": 328.304,
      "eval_steps_per_second": 5.165,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9778815507888794,
      "learning_rate": 4.666666666666666e-05,
      "loss": 0.0727,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9429333333333333,
      "eval_loss": 0.19682304561138153,
      "eval_runtime": 11.5774,
      "eval_samples_per_second": 323.907,
      "eval_steps_per_second": 5.096,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.0580928325653076,
      "learning_rate": 3.8888888888888884e-05,
      "loss": 0.0395,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9413333333333334,
      "eval_loss": 0.23067164421081543,
      "eval_runtime": 11.6038,
      "eval_samples_per_second": 323.171,
      "eval_steps_per_second": 5.085,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.5527564287185669,
      "learning_rate": 3.111111111111111e-05,
      "loss": 0.0215,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9429333333333333,
      "eval_loss": 0.22721877694129944,
      "eval_runtime": 12.3558,
      "eval_samples_per_second": 303.5,
      "eval_steps_per_second": 4.775,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.00954844057559967,
      "learning_rate": 2.333333333333333e-05,
      "loss": 0.0107,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9432,
      "eval_loss": 0.2322542816400528,
      "eval_runtime": 11.5419,
      "eval_samples_per_second": 324.902,
      "eval_steps_per_second": 5.112,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.557891224557891e-05,
      "loss": 0.0047,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9485333333333333,
      "eval_loss": 0.22237330675125122,
      "eval_runtime": 11.9525,
      "eval_samples_per_second": 313.741,
      "eval_steps_per_second": 4.936,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.011979671195149422,
      "learning_rate": 7.801134467801133e-06,
      "loss": 0.0022,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9469333333333333,
      "eval_loss": 0.22966329753398895,
      "eval_runtime": 11.4793,
      "eval_samples_per_second": 326.675,
      "eval_steps_per_second": 5.14,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.018474088981747627,
      "learning_rate": 2.3356690023356688e-08,
      "loss": 0.0017,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9496,
      "eval_loss": 0.2414996474981308,
      "eval_runtime": 11.5422,
      "eval_samples_per_second": 324.896,
      "eval_steps_per_second": 5.112,
      "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
}