trainer_state.json

{
  "best_metric": 0.9538666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0188/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": 0.25644397735595703,
      "learning_rate": 9.698463103929542e-05,
      "loss": 0.5705,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9325333333333333,
      "eval_loss": 0.24011661112308502,
      "eval_runtime": 8.726,
      "eval_samples_per_second": 429.751,
      "eval_steps_per_second": 6.761,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.05274973437190056,
      "learning_rate": 8.83022221559489e-05,
      "loss": 0.1435,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9352,
      "eval_loss": 0.22778242826461792,
      "eval_runtime": 7.8953,
      "eval_samples_per_second": 474.965,
      "eval_steps_per_second": 7.473,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.2534704208374023,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.072,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9352,
      "eval_loss": 0.22844859957695007,
      "eval_runtime": 7.4077,
      "eval_samples_per_second": 506.232,
      "eval_steps_per_second": 7.965,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.08621691167354584,
      "learning_rate": 5.868240888334653e-05,
      "loss": 0.0337,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9421333333333334,
      "eval_loss": 0.21702972054481506,
      "eval_runtime": 6.7875,
      "eval_samples_per_second": 552.487,
      "eval_steps_per_second": 8.692,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.04713824391365051,
      "learning_rate": 4.131759111665349e-05,
      "loss": 0.0139,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9474666666666667,
      "eval_loss": 0.22027580440044403,
      "eval_runtime": 7.1375,
      "eval_samples_per_second": 525.395,
      "eval_steps_per_second": 8.266,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.009096852503716946,
      "learning_rate": 2.500000000000001e-05,
      "loss": 0.0045,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9504,
      "eval_loss": 0.21194615960121155,
      "eval_runtime": 8.0113,
      "eval_samples_per_second": 468.087,
      "eval_steps_per_second": 7.365,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.1731488851999511e-05,
      "loss": 0.0019,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9509333333333333,
      "eval_loss": 0.20759093761444092,
      "eval_runtime": 7.6164,
      "eval_samples_per_second": 492.358,
      "eval_steps_per_second": 7.746,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.01624929904937744,
      "learning_rate": 3.033320830114261e-06,
      "loss": 0.0015,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9525333333333333,
      "eval_loss": 0.21065981686115265,
      "eval_runtime": 6.8601,
      "eval_samples_per_second": 546.641,
      "eval_steps_per_second": 8.6,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.031816910952329636,
      "learning_rate": 2.7470478758395702e-11,
      "loss": 0.001,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9538666666666666,
      "eval_loss": 0.21338486671447754,
      "eval_runtime": 7.2137,
      "eval_samples_per_second": 519.844,
      "eval_steps_per_second": 8.179,
      "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
}