trainer_state.json

{
  "best_metric": 0.42346666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0717/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": 0.00048527961022201525,
      "loss": 3.3654,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.19866666666666666,
      "eval_loss": 2.957345724105835,
      "eval_runtime": 13.5234,
      "eval_samples_per_second": 277.296,
      "eval_steps_per_second": 4.363,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 26.043676376342773,
      "learning_rate": 0.0004421832247293072,
      "loss": 2.9116,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2312,
      "eval_loss": 2.8191897869110107,
      "eval_runtime": 13.4193,
      "eval_samples_per_second": 279.449,
      "eval_steps_per_second": 4.397,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 21.010509490966797,
      "learning_rate": 0.00037590670601192445,
      "loss": 2.7308,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.27386666666666665,
      "eval_loss": 2.6436994075775146,
      "eval_runtime": 13.4948,
      "eval_samples_per_second": 277.885,
      "eval_steps_per_second": 4.372,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 24.506147384643555,
      "learning_rate": 0.00029444398036442514,
      "loss": 2.5465,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3216,
      "eval_loss": 2.48641300201416,
      "eval_runtime": 13.4347,
      "eval_samples_per_second": 279.127,
      "eval_steps_per_second": 4.392,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 30.12303924560547,
      "learning_rate": 0.00020762065476168375,
      "loss": 2.3938,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.332,
      "eval_loss": 2.4021530151367188,
      "eval_runtime": 13.4531,
      "eval_samples_per_second": 278.746,
      "eval_steps_per_second": 4.386,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 29.73305320739746,
      "learning_rate": 0.00012590890364720656,
      "loss": 2.2013,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.372,
      "eval_loss": 2.2626140117645264,
      "eval_runtime": 13.4726,
      "eval_samples_per_second": 278.342,
      "eval_steps_per_second": 4.379,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 16.639404296875,
      "learning_rate": 5.9164370142409884e-05,
      "loss": 2.0084,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4053333333333333,
      "eval_loss": 2.130749464035034,
      "eval_runtime": 13.2482,
      "eval_samples_per_second": 283.058,
      "eval_steps_per_second": 4.453,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 18.729055404663086,
      "learning_rate": 1.5437430032376602e-05,
      "loss": 1.8313,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.42213333333333336,
      "eval_loss": 2.1039528846740723,
      "eval_runtime": 13.541,
      "eval_samples_per_second": 276.936,
      "eval_steps_per_second": 4.357,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": Infinity,
      "learning_rate": 3.433802298419275e-09,
      "loss": 1.7374,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.42346666666666666,
      "eval_loss": 2.106276273727417,
      "eval_runtime": 13.2919,
      "eval_samples_per_second": 282.126,
      "eval_steps_per_second": 4.439,
      "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
}