trainer_state.json

{
  "best_metric": 0.8442666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0937/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": 1.3824213743209839,
      "learning_rate": 0.0002666666666666666,
      "loss": 1.6716,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.72,
      "eval_loss": 0.9346333146095276,
      "eval_runtime": 11.5867,
      "eval_samples_per_second": 323.646,
      "eval_steps_per_second": 5.092,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.00023343343343343342,
      "loss": 0.669,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7602666666666666,
      "eval_loss": 0.780525267124176,
      "eval_runtime": 10.2494,
      "eval_samples_per_second": 365.876,
      "eval_steps_per_second": 5.756,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": Infinity,
      "learning_rate": 0.00020020020020020016,
      "loss": 0.3705,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8034666666666667,
      "eval_loss": 0.6650471091270447,
      "eval_runtime": 10.2261,
      "eval_samples_per_second": 366.71,
      "eval_steps_per_second": 5.77,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.052090927958488464,
      "learning_rate": 0.00016686686686686686,
      "loss": 0.1954,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8048,
      "eval_loss": 0.7129369974136353,
      "eval_runtime": 10.3195,
      "eval_samples_per_second": 363.39,
      "eval_steps_per_second": 5.717,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 23.67034149169922,
      "learning_rate": 0.00013353353353353354,
      "loss": 0.0986,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8096,
      "eval_loss": 0.7377421855926514,
      "eval_runtime": 10.9336,
      "eval_samples_per_second": 342.979,
      "eval_steps_per_second": 5.396,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.013217899017035961,
      "learning_rate": 0.0001002002002002002,
      "loss": 0.048,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8301333333333333,
      "eval_loss": 0.6890658736228943,
      "eval_runtime": 11.2884,
      "eval_samples_per_second": 332.199,
      "eval_steps_per_second": 5.227,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.6103159785270691,
      "learning_rate": 6.686686686686686e-05,
      "loss": 0.0149,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8301333333333333,
      "eval_loss": 0.7243148684501648,
      "eval_runtime": 10.3036,
      "eval_samples_per_second": 363.949,
      "eval_steps_per_second": 5.726,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.013547549024224281,
      "learning_rate": 3.3533533533533535e-05,
      "loss": 0.0054,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8410666666666666,
      "eval_loss": 0.6951678395271301,
      "eval_runtime": 10.1167,
      "eval_samples_per_second": 370.676,
      "eval_steps_per_second": 5.832,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.20495545864105225,
      "learning_rate": 2.002002002002002e-07,
      "loss": 0.0019,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8442666666666667,
      "eval_loss": 0.6896828413009644,
      "eval_runtime": 10.0805,
      "eval_samples_per_second": 372.006,
      "eval_steps_per_second": 5.853,
      "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
}