trainer_state.json

{
  "best_metric": 0.8757333333333334,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0281/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": 172.0108642578125,
      "learning_rate": 4.752422169756048e-05,
      "loss": 3.557,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5994666666666667,
      "eval_loss": 2.7449934482574463,
      "eval_runtime": 7.2256,
      "eval_samples_per_second": 518.991,
      "eval_steps_per_second": 8.165,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 197.930419921875,
      "learning_rate": 4.058724504646834e-05,
      "loss": 1.46,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7944,
      "eval_loss": 0.8281282782554626,
      "eval_runtime": 7.1234,
      "eval_samples_per_second": 526.434,
      "eval_steps_per_second": 8.283,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 199.5099639892578,
      "learning_rate": 3.056302334890786e-05,
      "loss": 0.6611,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8482666666666666,
      "eval_loss": 0.5446193814277649,
      "eval_runtime": 7.4423,
      "eval_samples_per_second": 503.875,
      "eval_steps_per_second": 7.928,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 76.01134490966797,
      "learning_rate": 1.9436976651092144e-05,
      "loss": 0.493,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8605333333333334,
      "eval_loss": 0.47560814023017883,
      "eval_runtime": 7.3694,
      "eval_samples_per_second": 508.862,
      "eval_steps_per_second": 8.006,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 239.1002197265625,
      "learning_rate": 9.412754953531663e-06,
      "loss": 0.4159,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8682666666666666,
      "eval_loss": 0.4392426311969757,
      "eval_runtime": 7.2101,
      "eval_samples_per_second": 520.101,
      "eval_steps_per_second": 8.183,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 235.13568115234375,
      "learning_rate": 2.475778302439524e-06,
      "loss": 0.3834,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8717333333333334,
      "eval_loss": 0.432939350605011,
      "eval_runtime": 7.116,
      "eval_samples_per_second": 526.981,
      "eval_steps_per_second": 8.291,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 273.16265869140625,
      "learning_rate": 0.0,
      "loss": 0.3652,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8757333333333334,
      "eval_loss": 0.4148465394973755,
      "eval_runtime": 8.8121,
      "eval_samples_per_second": 425.551,
      "eval_steps_per_second": 6.695,
      "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": 5.7235004676864e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}