trainer_state.json

{
  "best_metric": 0.9552,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0314/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": 14.541894912719727,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.767,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9362666666666667,
      "eval_loss": 0.23523712158203125,
      "eval_runtime": 5.6858,
      "eval_samples_per_second": 659.537,
      "eval_steps_per_second": 10.377,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 22.461755752563477,
      "learning_rate": 3.888888888888889e-05,
      "loss": 0.1107,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9448,
      "eval_loss": 0.18821214139461517,
      "eval_runtime": 5.6559,
      "eval_samples_per_second": 663.028,
      "eval_steps_per_second": 10.432,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.0865599662065506,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.0385,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.948,
      "eval_loss": 0.1985563486814499,
      "eval_runtime": 5.9735,
      "eval_samples_per_second": 627.77,
      "eval_steps_per_second": 9.877,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.059078674763441086,
      "learning_rate": 2.777777777777778e-05,
      "loss": 0.0153,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9496,
      "eval_loss": 0.19788086414337158,
      "eval_runtime": 5.7343,
      "eval_samples_per_second": 653.954,
      "eval_steps_per_second": 10.289,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.020871704444289207,
      "learning_rate": 2.2222222222222223e-05,
      "loss": 0.0062,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9517333333333333,
      "eval_loss": 0.20122885704040527,
      "eval_runtime": 5.6594,
      "eval_samples_per_second": 662.609,
      "eval_steps_per_second": 10.425,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.01741022802889347,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0031,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9546666666666667,
      "eval_loss": 0.1973712295293808,
      "eval_runtime": 5.8131,
      "eval_samples_per_second": 645.091,
      "eval_steps_per_second": 10.149,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.1127794461127795e-05,
      "loss": 0.0015,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9536,
      "eval_loss": 0.20005081593990326,
      "eval_runtime": 5.7889,
      "eval_samples_per_second": 647.794,
      "eval_steps_per_second": 10.192,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.016990995034575462,
      "learning_rate": 5.572238905572239e-06,
      "loss": 0.0015,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9549333333333333,
      "eval_loss": 0.20384365320205688,
      "eval_runtime": 5.9747,
      "eval_samples_per_second": 627.651,
      "eval_steps_per_second": 9.875,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.021080419421195984,
      "learning_rate": 1.6683350016683353e-08,
      "loss": 0.001,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9552,
      "eval_loss": 0.1995849609375,
      "eval_runtime": 5.9222,
      "eval_samples_per_second": 633.211,
      "eval_steps_per_second": 9.963,
      "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
}