trainer_state.json

{
  "best_metric": 0.9394666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0005/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": 14.145313262939453,
      "learning_rate": 6.788924172750679e-05,
      "loss": 0.7738,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9146666666666666,
      "eval_loss": 0.2999880909919739,
      "eval_runtime": 13.714,
      "eval_samples_per_second": 273.443,
      "eval_steps_per_second": 4.302,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.22387737035751343,
      "learning_rate": 6.181155550916422e-05,
      "loss": 0.1656,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.916,
      "eval_loss": 0.27056923508644104,
      "eval_runtime": 12.7058,
      "eval_samples_per_second": 295.141,
      "eval_steps_per_second": 4.644,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.36915922164917,
      "learning_rate": 5.2499999999999995e-05,
      "loss": 0.0799,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9242666666666667,
      "eval_loss": 0.2807912230491638,
      "eval_runtime": 13.4028,
      "eval_samples_per_second": 279.792,
      "eval_steps_per_second": 4.402,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.03225693479180336,
      "learning_rate": 4.1077686218342566e-05,
      "loss": 0.0428,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9285333333333333,
      "eval_loss": 0.2675916254520416,
      "eval_runtime": 13.3239,
      "eval_samples_per_second": 281.449,
      "eval_steps_per_second": 4.428,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.009314323775470257,
      "learning_rate": 2.8922313781657438e-05,
      "loss": 0.0172,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9349333333333333,
      "eval_loss": 0.27047550678253174,
      "eval_runtime": 13.4436,
      "eval_samples_per_second": 278.943,
      "eval_steps_per_second": 4.389,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.05501106381416321,
      "learning_rate": 1.7500000000000005e-05,
      "loss": 0.0105,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9365333333333333,
      "eval_loss": 0.26677557826042175,
      "eval_runtime": 13.1543,
      "eval_samples_per_second": 285.078,
      "eval_steps_per_second": 4.485,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 8.212042196399656e-06,
      "loss": 0.0055,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9381333333333334,
      "eval_loss": 0.26110953092575073,
      "eval_runtime": 13.2532,
      "eval_samples_per_second": 282.95,
      "eval_steps_per_second": 4.452,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.005280699580907822,
      "learning_rate": 2.1233245810799825e-06,
      "loss": 0.0037,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9349333333333333,
      "eval_loss": 0.27361005544662476,
      "eval_runtime": 13.2842,
      "eval_samples_per_second": 282.291,
      "eval_steps_per_second": 4.441,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.04525449499487877,
      "learning_rate": 1.9229335130876987e-11,
      "loss": 0.0028,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9394666666666667,
      "eval_loss": 0.27074283361434937,
      "eval_runtime": 13.4084,
      "eval_samples_per_second": 279.675,
      "eval_steps_per_second": 4.4,
      "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
}