trainer_state.json

{
  "best_metric": 0.9064,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0632/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": 67.94400787353516,
      "learning_rate": 6.125e-05,
      "loss": 3.3409,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6709333333333334,
      "eval_loss": 1.7803062200546265,
      "eval_runtime": 8.8423,
      "eval_samples_per_second": 424.098,
      "eval_steps_per_second": 6.672,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 74.40616607666016,
      "learning_rate": 5.2499999999999995e-05,
      "loss": 0.8968,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8546666666666667,
      "eval_loss": 0.5411354899406433,
      "eval_runtime": 8.2032,
      "eval_samples_per_second": 457.141,
      "eval_steps_per_second": 7.192,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 212.97666931152344,
      "learning_rate": 4.374999999999999e-05,
      "loss": 0.4594,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8834666666666666,
      "eval_loss": 0.4071122705936432,
      "eval_runtime": 8.0797,
      "eval_samples_per_second": 464.128,
      "eval_steps_per_second": 7.302,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 77.5980224609375,
      "learning_rate": 3.5e-05,
      "loss": 0.3303,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8952,
      "eval_loss": 0.35080018639564514,
      "eval_runtime": 8.2229,
      "eval_samples_per_second": 456.042,
      "eval_steps_per_second": 7.175,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 58.72169876098633,
      "learning_rate": 2.6249999999999998e-05,
      "loss": 0.251,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.896,
      "eval_loss": 0.33999302983283997,
      "eval_runtime": 8.4857,
      "eval_samples_per_second": 441.92,
      "eval_steps_per_second": 6.953,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 353.4154052734375,
      "learning_rate": 1.75e-05,
      "loss": 0.2079,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9029333333333334,
      "eval_loss": 0.32216787338256836,
      "eval_runtime": 8.1992,
      "eval_samples_per_second": 457.362,
      "eval_steps_per_second": 7.196,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 128.898681640625,
      "learning_rate": 8.75e-06,
      "loss": 0.1912,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9064,
      "eval_loss": 0.3101765513420105,
      "eval_runtime": 8.6134,
      "eval_samples_per_second": 435.366,
      "eval_steps_per_second": 6.85,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2664,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.7235004676864e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}