trainer_state.json

{
  "best_metric": 0.9210666666666667,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0098/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": 18.44796371459961,
      "learning_rate": 0.00042857142857142855,
      "loss": 1.3278,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8525333333333334,
      "eval_loss": 0.5378308892250061,
      "eval_runtime": 8.124,
      "eval_samples_per_second": 461.593,
      "eval_steps_per_second": 7.262,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 41.9058723449707,
      "learning_rate": 0.00035714285714285714,
      "loss": 0.3945,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8888,
      "eval_loss": 0.40475594997406006,
      "eval_runtime": 7.712,
      "eval_samples_per_second": 486.254,
      "eval_steps_per_second": 7.65,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 43.523250579833984,
      "learning_rate": 0.0002857142857142857,
      "loss": 0.2157,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8957333333333334,
      "eval_loss": 0.3760475516319275,
      "eval_runtime": 7.6787,
      "eval_samples_per_second": 488.361,
      "eval_steps_per_second": 7.684,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 22.369220733642578,
      "learning_rate": 0.00021428571428571427,
      "loss": 0.1331,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9061333333333333,
      "eval_loss": 0.35818254947662354,
      "eval_runtime": 7.6901,
      "eval_samples_per_second": 487.64,
      "eval_steps_per_second": 7.672,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 23.01726722717285,
      "learning_rate": 0.00014285714285714284,
      "loss": 0.0701,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.912,
      "eval_loss": 0.38542264699935913,
      "eval_runtime": 8.3049,
      "eval_samples_per_second": 451.54,
      "eval_steps_per_second": 7.104,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 40.775794982910156,
      "learning_rate": 7.142857142857142e-05,
      "loss": 0.0461,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.912,
      "eval_loss": 0.38149434328079224,
      "eval_runtime": 7.5877,
      "eval_samples_per_second": 494.218,
      "eval_steps_per_second": 7.776,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 34.34041976928711,
      "learning_rate": 0.0,
      "loss": 0.0342,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9210666666666667,
      "eval_loss": 0.3559032678604126,
      "eval_runtime": 8.0436,
      "eval_samples_per_second": 466.208,
      "eval_steps_per_second": 7.335,
      "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
}