trainer_state.json

{
  "best_metric": 0.8736,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0852/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": Infinity,
      "learning_rate": 2.667667667667668e-05,
      "loss": 2.4553,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6797333333333333,
      "eval_loss": 1.3701738119125366,
      "eval_runtime": 13.3466,
      "eval_samples_per_second": 280.97,
      "eval_steps_per_second": 4.421,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 8.25568675994873,
      "learning_rate": 2.3343343343343343e-05,
      "loss": 0.9321,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7792,
      "eval_loss": 0.8259291648864746,
      "eval_runtime": 13.011,
      "eval_samples_per_second": 288.217,
      "eval_steps_per_second": 4.535,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.6318492889404297,
      "learning_rate": 2.001001001001001e-05,
      "loss": 0.5429,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8317333333333333,
      "eval_loss": 0.6083554625511169,
      "eval_runtime": 14.0034,
      "eval_samples_per_second": 267.791,
      "eval_steps_per_second": 4.213,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9618034958839417,
      "learning_rate": 1.6676676676676678e-05,
      "loss": 0.3593,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8349333333333333,
      "eval_loss": 0.5686675906181335,
      "eval_runtime": 12.998,
      "eval_samples_per_second": 288.506,
      "eval_steps_per_second": 4.539,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 14.219752311706543,
      "learning_rate": 1.3343343343343344e-05,
      "loss": 0.2523,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8552,
      "eval_loss": 0.4929419159889221,
      "eval_runtime": 13.8706,
      "eval_samples_per_second": 270.356,
      "eval_steps_per_second": 4.254,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.7723649740219116,
      "learning_rate": 1.001001001001001e-05,
      "loss": 0.1737,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8616,
      "eval_loss": 0.4878195822238922,
      "eval_runtime": 13.0989,
      "eval_samples_per_second": 286.283,
      "eval_steps_per_second": 4.504,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": Infinity,
      "learning_rate": 6.686686686686687e-06,
      "loss": 0.1201,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8666666666666667,
      "eval_loss": 0.46121037006378174,
      "eval_runtime": 13.3781,
      "eval_samples_per_second": 280.31,
      "eval_steps_per_second": 4.41,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 4.142410755157471,
      "learning_rate": 3.353353353353354e-06,
      "loss": 0.0811,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8706666666666667,
      "eval_loss": 0.4670596420764923,
      "eval_runtime": 13.3756,
      "eval_samples_per_second": 280.362,
      "eval_steps_per_second": 4.411,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 10.538165092468262,
      "learning_rate": 2.002002002002002e-08,
      "loss": 0.0606,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8736,
      "eval_loss": 0.4490545988082886,
      "eval_runtime": 13.2283,
      "eval_samples_per_second": 283.483,
      "eval_steps_per_second": 4.46,
      "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
}