trainer_state.json

{
  "best_metric": 0.9216,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0625/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": 0.2246047407388687,
      "learning_rate": 0.0002909538931178862,
      "loss": 0.6281,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8669333333333333,
      "eval_loss": 0.42683476209640503,
      "eval_runtime": 7.8123,
      "eval_samples_per_second": 480.01,
      "eval_steps_per_second": 7.552,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.753866195678711,
      "learning_rate": 0.00026490666646784665,
      "loss": 0.2911,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8661333333333333,
      "eval_loss": 0.468487411737442,
      "eval_runtime": 7.5625,
      "eval_samples_per_second": 495.867,
      "eval_steps_per_second": 7.802,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.1162219047546387,
      "learning_rate": 0.000225,
      "loss": 0.1797,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8909333333333334,
      "eval_loss": 0.3809038996696472,
      "eval_runtime": 7.4513,
      "eval_samples_per_second": 503.267,
      "eval_steps_per_second": 7.918,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.41836875677108765,
      "learning_rate": 0.00017604722665003956,
      "loss": 0.108,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8773333333333333,
      "eval_loss": 0.4885323941707611,
      "eval_runtime": 7.9653,
      "eval_samples_per_second": 470.791,
      "eval_steps_per_second": 7.407,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.02262718789279461,
      "learning_rate": 0.00012395277334996044,
      "loss": 0.0547,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8946666666666667,
      "eval_loss": 0.4177989959716797,
      "eval_runtime": 7.2739,
      "eval_samples_per_second": 515.54,
      "eval_steps_per_second": 8.111,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.003164676483720541,
      "learning_rate": 7.500000000000002e-05,
      "loss": 0.0215,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9125333333333333,
      "eval_loss": 0.35623010993003845,
      "eval_runtime": 7.3674,
      "eval_samples_per_second": 508.997,
      "eval_steps_per_second": 8.008,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 3.519446655599853e-05,
      "loss": 0.0065,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9178666666666667,
      "eval_loss": 0.3263644278049469,
      "eval_runtime": 7.6995,
      "eval_samples_per_second": 487.042,
      "eval_steps_per_second": 7.663,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.007696530316025019,
      "learning_rate": 9.099962490342783e-06,
      "loss": 0.0029,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9202666666666667,
      "eval_loss": 0.32276806235313416,
      "eval_runtime": 7.5572,
      "eval_samples_per_second": 496.215,
      "eval_steps_per_second": 7.807,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.03857341781258583,
      "learning_rate": 8.24114362751871e-11,
      "loss": 0.0016,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9216,
      "eval_loss": 0.3386729061603546,
      "eval_runtime": 7.6668,
      "eval_samples_per_second": 489.119,
      "eval_steps_per_second": 7.695,
      "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
}