trainer_state.json

{
  "best_metric": 0.9384,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0437/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.495139718055725,
      "learning_rate": 7.714285714285714e-05,
      "loss": 0.6355,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9197333333333333,
      "eval_loss": 0.26206839084625244,
      "eval_runtime": 6.6032,
      "eval_samples_per_second": 567.905,
      "eval_steps_per_second": 8.935,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.052195172756910324,
      "learning_rate": 6.428571428571429e-05,
      "loss": 0.1293,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9210666666666667,
      "eval_loss": 0.26283159852027893,
      "eval_runtime": 6.2823,
      "eval_samples_per_second": 596.912,
      "eval_steps_per_second": 9.391,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.04223431646823883,
      "learning_rate": 5.142857142857143e-05,
      "loss": 0.0507,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9274666666666667,
      "eval_loss": 0.2899295687675476,
      "eval_runtime": 5.605,
      "eval_samples_per_second": 669.049,
      "eval_steps_per_second": 10.526,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.015206639654934406,
      "learning_rate": 3.857142857142857e-05,
      "loss": 0.0199,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9333333333333333,
      "eval_loss": 0.2675316631793976,
      "eval_runtime": 5.838,
      "eval_samples_per_second": 642.339,
      "eval_steps_per_second": 10.106,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.01165010966360569,
      "learning_rate": 2.5714285714285714e-05,
      "loss": 0.0077,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9354666666666667,
      "eval_loss": 0.27498576045036316,
      "eval_runtime": 5.8827,
      "eval_samples_per_second": 637.465,
      "eval_steps_per_second": 10.029,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.009392859414219856,
      "learning_rate": 1.2857142857142857e-05,
      "loss": 0.0035,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.936,
      "eval_loss": 0.27553924918174744,
      "eval_runtime": 5.842,
      "eval_samples_per_second": 641.907,
      "eval_steps_per_second": 10.099,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 3.861003861003861e-08,
      "loss": 0.0018,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9384,
      "eval_loss": 0.26887017488479614,
      "eval_runtime": 5.8531,
      "eval_samples_per_second": 640.69,
      "eval_steps_per_second": 10.08,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}