trainer_state.json

{
  "best_metric": 0.4146666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0215/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": 21.073204040527344,
      "learning_rate": 0.0003,
      "loss": 3.3536,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.208,
      "eval_loss": 3.014976978302002,
      "eval_runtime": 7.5956,
      "eval_samples_per_second": 493.708,
      "eval_steps_per_second": 7.768,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 27.168363571166992,
      "learning_rate": 0.0003,
      "loss": 2.8447,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.24346666666666666,
      "eval_loss": 2.8262526988983154,
      "eval_runtime": 6.88,
      "eval_samples_per_second": 545.061,
      "eval_steps_per_second": 8.576,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 20.78453826904297,
      "learning_rate": 0.0003,
      "loss": 2.6447,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.27813333333333334,
      "eval_loss": 2.673530101776123,
      "eval_runtime": 6.8216,
      "eval_samples_per_second": 549.724,
      "eval_steps_per_second": 8.649,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 22.504783630371094,
      "learning_rate": 0.0003,
      "loss": 2.4929,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.29706666666666665,
      "eval_loss": 2.556272506713867,
      "eval_runtime": 7.7848,
      "eval_samples_per_second": 481.711,
      "eval_steps_per_second": 7.579,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 20.437288284301758,
      "learning_rate": 0.0003,
      "loss": 2.3454,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.32986666666666664,
      "eval_loss": 2.4549481868743896,
      "eval_runtime": 7.9301,
      "eval_samples_per_second": 472.884,
      "eval_steps_per_second": 7.44,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 19.95916175842285,
      "learning_rate": 0.0003,
      "loss": 2.2138,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.36666666666666664,
      "eval_loss": 2.3166587352752686,
      "eval_runtime": 7.3728,
      "eval_samples_per_second": 508.627,
      "eval_steps_per_second": 8.002,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0003,
      "loss": 2.0954,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.38826666666666665,
      "eval_loss": 2.21130108833313,
      "eval_runtime": 6.8763,
      "eval_samples_per_second": 545.354,
      "eval_steps_per_second": 8.58,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 18.77606201171875,
      "learning_rate": 0.0003,
      "loss": 1.9542,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.3792,
      "eval_loss": 2.2539145946502686,
      "eval_runtime": 6.7781,
      "eval_samples_per_second": 553.256,
      "eval_steps_per_second": 8.705,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 23.879175186157227,
      "learning_rate": 0.0003,
      "loss": 1.8381,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4146666666666667,
      "eval_loss": 2.1466076374053955,
      "eval_runtime": 6.8199,
      "eval_samples_per_second": 549.864,
      "eval_steps_per_second": 8.651,
      "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
}