trainer_state.json

{
  "best_metric": 0.9168,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0820/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.009199619293213,
      "learning_rate": 0.0002851453301853628,
      "loss": 0.6958,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8450666666666666,
      "eval_loss": 0.5190665125846863,
      "eval_runtime": 5.8496,
      "eval_samples_per_second": 641.065,
      "eval_steps_per_second": 10.086,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2604600191116333,
      "learning_rate": 0.00024352347027881003,
      "loss": 0.3272,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8693333333333333,
      "eval_loss": 0.43013885617256165,
      "eval_runtime": 5.7914,
      "eval_samples_per_second": 647.517,
      "eval_steps_per_second": 10.188,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.022271357476711273,
      "learning_rate": 0.00018337814009344714,
      "loss": 0.1867,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8786666666666667,
      "eval_loss": 0.4129250943660736,
      "eval_runtime": 5.8579,
      "eval_samples_per_second": 640.158,
      "eval_steps_per_second": 10.072,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3810561895370483,
      "learning_rate": 0.00011662185990655284,
      "loss": 0.0994,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8938666666666667,
      "eval_loss": 0.4062184989452362,
      "eval_runtime": 5.9535,
      "eval_samples_per_second": 629.886,
      "eval_steps_per_second": 9.91,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.011195636354386806,
      "learning_rate": 5.6476529721189974e-05,
      "loss": 0.0343,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9104,
      "eval_loss": 0.3616520166397095,
      "eval_runtime": 5.8027,
      "eval_samples_per_second": 646.25,
      "eval_steps_per_second": 10.168,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.004247686825692654,
      "learning_rate": 1.4854669814637143e-05,
      "loss": 0.01,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9128,
      "eval_loss": 0.33179861307144165,
      "eval_runtime": 5.8565,
      "eval_samples_per_second": 640.318,
      "eval_steps_per_second": 10.074,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.3623114161753767e-10,
      "loss": 0.0047,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9168,
      "eval_loss": 0.33798593282699585,
      "eval_runtime": 5.9634,
      "eval_samples_per_second": 628.834,
      "eval_steps_per_second": 9.894,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}