trainer_state.json

{
  "best_metric": 0.88,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0853/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": 124.16468811035156,
      "learning_rate": 9e-05,
      "loss": 2.8771,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7184,
      "eval_loss": 1.0233898162841797,
      "eval_runtime": 8.6809,
      "eval_samples_per_second": 431.984,
      "eval_steps_per_second": 6.797,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 83.75138092041016,
      "learning_rate": 9e-05,
      "loss": 0.7443,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8288,
      "eval_loss": 0.5401186347007751,
      "eval_runtime": 8.881,
      "eval_samples_per_second": 422.251,
      "eval_steps_per_second": 6.643,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 83.9761962890625,
      "learning_rate": 9e-05,
      "loss": 0.4601,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8525333333333334,
      "eval_loss": 0.46282896399497986,
      "eval_runtime": 8.7326,
      "eval_samples_per_second": 429.425,
      "eval_steps_per_second": 6.756,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 44.836265563964844,
      "learning_rate": 9e-05,
      "loss": 0.3328,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8696,
      "eval_loss": 0.41567620635032654,
      "eval_runtime": 9.0032,
      "eval_samples_per_second": 416.518,
      "eval_steps_per_second": 6.553,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 41.71147537231445,
      "learning_rate": 9e-05,
      "loss": 0.2493,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8706666666666667,
      "eval_loss": 0.4032222628593445,
      "eval_runtime": 8.619,
      "eval_samples_per_second": 435.084,
      "eval_steps_per_second": 6.845,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 91.97679138183594,
      "learning_rate": 9e-05,
      "loss": 0.1876,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8754666666666666,
      "eval_loss": 0.40024518966674805,
      "eval_runtime": 8.5659,
      "eval_samples_per_second": 437.782,
      "eval_steps_per_second": 6.888,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 81.4844970703125,
      "learning_rate": 9e-05,
      "loss": 0.1505,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.88,
      "eval_loss": 0.41271263360977173,
      "eval_runtime": 8.6,
      "eval_samples_per_second": 436.049,
      "eval_steps_per_second": 6.861,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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
}