trainer_state.json

{
  "best_metric": 0.9317333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_1001/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": 6.7914265912926e-05,
      "loss": 0.7328,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8778666666666667,
      "eval_loss": 0.3821270763874054,
      "eval_runtime": 10.3565,
      "eval_samples_per_second": 362.092,
      "eval_steps_per_second": 5.697,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 37.513526916503906,
      "learning_rate": 6.185866251124284e-05,
      "loss": 0.2211,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8725333333333334,
      "eval_loss": 0.44321829080581665,
      "eval_runtime": 10.5982,
      "eval_samples_per_second": 353.834,
      "eval_steps_per_second": 5.567,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.0801468938589096,
      "learning_rate": 5.2563508019062023e-05,
      "loss": 0.1185,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8802666666666666,
      "eval_loss": 0.4580089747905731,
      "eval_runtime": 10.7123,
      "eval_samples_per_second": 350.063,
      "eval_steps_per_second": 5.508,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.0057269977405667305,
      "learning_rate": 4.1149935250010576e-05,
      "loss": 0.0577,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9088,
      "eval_loss": 0.38068145513534546,
      "eval_runtime": 10.3766,
      "eval_samples_per_second": 361.391,
      "eval_steps_per_second": 5.686,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.007332438137382269,
      "learning_rate": 2.8994589526430138e-05,
      "loss": 0.0335,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9010666666666667,
      "eval_loss": 0.40658804774284363,
      "eval_runtime": 10.8835,
      "eval_samples_per_second": 344.557,
      "eval_steps_per_second": 5.421,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.005678305868059397,
      "learning_rate": 1.756358493638142e-05,
      "loss": 0.0077,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9213333333333333,
      "eval_loss": 0.338676780462265,
      "eval_runtime": 11.1848,
      "eval_samples_per_second": 335.278,
      "eval_steps_per_second": 5.275,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 2.5623437977628782e-05,
      "learning_rate": 8.235669337084578e-06,
      "loss": 0.0025,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9261333333333334,
      "eval_loss": 0.32309582829475403,
      "eval_runtime": 10.4033,
      "eval_samples_per_second": 360.462,
      "eval_steps_per_second": 5.671,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.002755251945927739,
      "learning_rate": 2.135927015187301e-06,
      "loss": 0.001,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9288,
      "eval_loss": 0.30905747413635254,
      "eval_runtime": 10.3963,
      "eval_samples_per_second": 360.705,
      "eval_steps_per_second": 5.675,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 5.346420584828593e-05,
      "learning_rate": 7.691731939263313e-11,
      "loss": 0.0003,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9317333333333333,
      "eval_loss": 0.3021787106990814,
      "eval_runtime": 10.5409,
      "eval_samples_per_second": 355.758,
      "eval_steps_per_second": 5.597,
      "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
}