trainer_state.json

{
  "best_metric": 0.9456,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0651/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": 0.865867555141449,
      "learning_rate": 7.714285714285714e-05,
      "loss": 0.6986,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9138666666666667,
      "eval_loss": 0.28853657841682434,
      "eval_runtime": 13.4197,
      "eval_samples_per_second": 279.441,
      "eval_steps_per_second": 4.397,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.15929360687732697,
      "learning_rate": 6.428571428571429e-05,
      "loss": 0.1556,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9242666666666667,
      "eval_loss": 0.25232625007629395,
      "eval_runtime": 13.5307,
      "eval_samples_per_second": 277.148,
      "eval_steps_per_second": 4.36,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.2195921391248703,
      "learning_rate": 5.142857142857143e-05,
      "loss": 0.0734,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9354666666666667,
      "eval_loss": 0.23003575205802917,
      "eval_runtime": 12.3576,
      "eval_samples_per_second": 303.458,
      "eval_steps_per_second": 4.774,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.06326346844434738,
      "learning_rate": 3.857142857142857e-05,
      "loss": 0.0404,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9373333333333334,
      "eval_loss": 0.23802228271961212,
      "eval_runtime": 13.483,
      "eval_samples_per_second": 278.128,
      "eval_steps_per_second": 4.376,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.018446672707796097,
      "learning_rate": 2.5714285714285714e-05,
      "loss": 0.0183,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9413333333333334,
      "eval_loss": 0.22661468386650085,
      "eval_runtime": 14.2124,
      "eval_samples_per_second": 263.855,
      "eval_steps_per_second": 4.151,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 13.553123474121094,
      "learning_rate": 1.2857142857142857e-05,
      "loss": 0.0099,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9448,
      "eval_loss": 0.22276875376701355,
      "eval_runtime": 12.8057,
      "eval_samples_per_second": 292.838,
      "eval_steps_per_second": 4.607,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 3.861003861003861e-08,
      "loss": 0.0047,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9456,
      "eval_loss": 0.217005655169487,
      "eval_runtime": 13.4988,
      "eval_samples_per_second": 277.803,
      "eval_steps_per_second": 4.371,
      "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
}