trainer_state.json

{
  "best_metric": 0.8552,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0013/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": 139.20338439941406,
      "learning_rate": 4.375e-05,
      "loss": 3.7611,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4189333333333333,
      "eval_loss": 3.3926267623901367,
      "eval_runtime": 9.0797,
      "eval_samples_per_second": 413.008,
      "eval_steps_per_second": 6.498,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 50.49544906616211,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 2.2459,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6930666666666667,
      "eval_loss": 1.3002595901489258,
      "eval_runtime": 8.8093,
      "eval_samples_per_second": 425.685,
      "eval_steps_per_second": 6.697,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 310.1706848144531,
      "learning_rate": 3.125e-05,
      "loss": 1.0225,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7909333333333334,
      "eval_loss": 0.7798284888267517,
      "eval_runtime": 8.8645,
      "eval_samples_per_second": 423.037,
      "eval_steps_per_second": 6.656,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 130.19073486328125,
      "learning_rate": 2.5e-05,
      "loss": 0.7322,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8090666666666667,
      "eval_loss": 0.6743772029876709,
      "eval_runtime": 8.9194,
      "eval_samples_per_second": 420.434,
      "eval_steps_per_second": 6.615,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 98.03427124023438,
      "learning_rate": 1.8750000000000002e-05,
      "loss": 0.6065,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.84,
      "eval_loss": 0.5687465071678162,
      "eval_runtime": 9.1321,
      "eval_samples_per_second": 410.64,
      "eval_steps_per_second": 6.461,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 207.90524291992188,
      "learning_rate": 1.25e-05,
      "loss": 0.545,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8477333333333333,
      "eval_loss": 0.5428369045257568,
      "eval_runtime": 8.7867,
      "eval_samples_per_second": 426.781,
      "eval_steps_per_second": 6.715,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 120.5766372680664,
      "learning_rate": 6.25e-06,
      "loss": 0.5103,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8552,
      "eval_loss": 0.5077586770057678,
      "eval_runtime": 8.7023,
      "eval_samples_per_second": 430.92,
      "eval_steps_per_second": 6.78,
      "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
}