trainer_state.json

{
  "best_metric": 0.8704,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0440/checkpoints/checkpoint-1998",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1998,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 21.681320190429688,
      "learning_rate": 7e-05,
      "loss": 0.8196,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8493333333333334,
      "eval_loss": 0.5102661848068237,
      "eval_runtime": 6.6524,
      "eval_samples_per_second": 563.705,
      "eval_steps_per_second": 8.869,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 30.040678024291992,
      "learning_rate": 7e-05,
      "loss": 0.3015,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8648,
      "eval_loss": 0.4530234932899475,
      "eval_runtime": 6.7401,
      "eval_samples_per_second": 556.369,
      "eval_steps_per_second": 8.754,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.22259506583213806,
      "learning_rate": 7e-05,
      "loss": 0.1835,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.86,
      "eval_loss": 0.512382447719574,
      "eval_runtime": 6.686,
      "eval_samples_per_second": 560.875,
      "eval_steps_per_second": 8.824,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.3172606229782104,
      "learning_rate": 7e-05,
      "loss": 0.1368,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8584,
      "eval_loss": 0.5354259014129639,
      "eval_runtime": 6.5551,
      "eval_samples_per_second": 572.075,
      "eval_steps_per_second": 9.001,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": Infinity,
      "learning_rate": 7e-05,
      "loss": 0.1104,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8597333333333333,
      "eval_loss": 0.5666179656982422,
      "eval_runtime": 6.9951,
      "eval_samples_per_second": 536.092,
      "eval_steps_per_second": 8.435,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.000722456315997988,
      "learning_rate": 7e-05,
      "loss": 0.1005,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8704,
      "eval_loss": 0.5145488977432251,
      "eval_runtime": 6.7062,
      "eval_samples_per_second": 559.186,
      "eval_steps_per_second": 8.798,
      "step": 1998
    }
  ],
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 9.88447923735552e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}