trainer_state.json

{
  "best_metric": 0.9578666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0391/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": 8.962423324584961,
      "learning_rate": 8.554359905560886e-05,
      "loss": 0.5838,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9368,
      "eval_loss": 0.2039853036403656,
      "eval_runtime": 5.2619,
      "eval_samples_per_second": 712.668,
      "eval_steps_per_second": 11.213,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.1351693719625473,
      "learning_rate": 7.305704108364301e-05,
      "loss": 0.0998,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9426666666666667,
      "eval_loss": 0.20324566960334778,
      "eval_runtime": 5.257,
      "eval_samples_per_second": 713.332,
      "eval_steps_per_second": 11.223,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.020450051873922348,
      "learning_rate": 5.501344202803415e-05,
      "loss": 0.0337,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9490666666666666,
      "eval_loss": 0.1939111053943634,
      "eval_runtime": 5.4153,
      "eval_samples_per_second": 692.481,
      "eval_steps_per_second": 10.895,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.02479933761060238,
      "learning_rate": 3.498655797196586e-05,
      "loss": 0.0108,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9544,
      "eval_loss": 0.1921376734972,
      "eval_runtime": 5.2696,
      "eval_samples_per_second": 711.627,
      "eval_steps_per_second": 11.196,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.2659974694252014,
      "learning_rate": 1.6942958916356995e-05,
      "loss": 0.0028,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.956,
      "eval_loss": 0.18728932738304138,
      "eval_runtime": 5.2087,
      "eval_samples_per_second": 719.953,
      "eval_steps_per_second": 11.327,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.013920309953391552,
      "learning_rate": 4.456400944391144e-06,
      "loss": 0.0014,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9576,
      "eval_loss": 0.1781616359949112,
      "eval_runtime": 5.647,
      "eval_samples_per_second": 664.075,
      "eval_steps_per_second": 10.448,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 4.086934248526131e-11,
      "loss": 0.0011,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9578666666666666,
      "eval_loss": 0.17801839113235474,
      "eval_runtime": 7.2654,
      "eval_samples_per_second": 516.148,
      "eval_steps_per_second": 8.121,
      "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
}