trainer_state.json

{
  "best_metric": 0.4093333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0193/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": 19.111595153808594,
      "learning_rate": 0.00044494494494494495,
      "loss": 3.512,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.18533333333333332,
      "eval_loss": 3.094391107559204,
      "eval_runtime": 7.0406,
      "eval_samples_per_second": 532.628,
      "eval_steps_per_second": 8.38,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 18.063495635986328,
      "learning_rate": 0.0003893893893893894,
      "loss": 2.9952,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2264,
      "eval_loss": 2.921708345413208,
      "eval_runtime": 6.96,
      "eval_samples_per_second": 538.792,
      "eval_steps_per_second": 8.477,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 19.978370666503906,
      "learning_rate": 0.00033383383383383386,
      "loss": 2.7713,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.2826666666666667,
      "eval_loss": 2.7162060737609863,
      "eval_runtime": 7.1107,
      "eval_samples_per_second": 527.375,
      "eval_steps_per_second": 8.297,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 31.8958740234375,
      "learning_rate": 0.00027827827827827826,
      "loss": 2.5684,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.29786666666666667,
      "eval_loss": 2.568107843399048,
      "eval_runtime": 7.1996,
      "eval_samples_per_second": 520.861,
      "eval_steps_per_second": 8.195,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 11.451271057128906,
      "learning_rate": 0.00022272272272272272,
      "loss": 2.3886,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.3384,
      "eval_loss": 2.436866044998169,
      "eval_runtime": 7.0903,
      "eval_samples_per_second": 528.895,
      "eval_steps_per_second": 8.321,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 26.946224212646484,
      "learning_rate": 0.00016716716716716717,
      "loss": 2.1928,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.3698666666666667,
      "eval_loss": 2.348764181137085,
      "eval_runtime": 7.1377,
      "eval_samples_per_second": 525.38,
      "eval_steps_per_second": 8.266,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": Infinity,
      "learning_rate": 0.00011177844511177845,
      "loss": 2.0047,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.3858666666666667,
      "eval_loss": 2.25692081451416,
      "eval_runtime": 7.123,
      "eval_samples_per_second": 526.46,
      "eval_steps_per_second": 8.283,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 39.8272590637207,
      "learning_rate": 5.622288955622289e-05,
      "loss": 1.7665,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.3992,
      "eval_loss": 2.2510833740234375,
      "eval_runtime": 7.2722,
      "eval_samples_per_second": 515.665,
      "eval_steps_per_second": 8.113,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 19.951377868652344,
      "learning_rate": 6.67334000667334e-07,
      "loss": 1.492,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4093333333333333,
      "eval_loss": 2.3158013820648193,
      "eval_runtime": 8.6099,
      "eval_samples_per_second": 435.544,
      "eval_steps_per_second": 6.853,
      "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
}