trainer_state.json

{
  "best_metric": 0.8608,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0624/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": 146.04156494140625,
      "learning_rate": 9.504844339512095e-05,
      "loss": 2.8445,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7221333333333333,
      "eval_loss": 1.0253187417984009,
      "eval_runtime": 13.7631,
      "eval_samples_per_second": 272.468,
      "eval_steps_per_second": 4.287,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 120.62777709960938,
      "learning_rate": 8.117449009293668e-05,
      "loss": 0.7717,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8181333333333334,
      "eval_loss": 0.5907644629478455,
      "eval_runtime": 13.3181,
      "eval_samples_per_second": 281.571,
      "eval_steps_per_second": 4.43,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 93.74173736572266,
      "learning_rate": 6.112604669781572e-05,
      "loss": 0.5007,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8378666666666666,
      "eval_loss": 0.5052756071090698,
      "eval_runtime": 13.2585,
      "eval_samples_per_second": 282.838,
      "eval_steps_per_second": 4.45,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 74.6573257446289,
      "learning_rate": 3.887395330218429e-05,
      "loss": 0.3673,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8522666666666666,
      "eval_loss": 0.45209959149360657,
      "eval_runtime": 13.2821,
      "eval_samples_per_second": 282.336,
      "eval_steps_per_second": 4.442,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 51.36497116088867,
      "learning_rate": 1.8825509907063327e-05,
      "loss": 0.3031,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4493885338306427,
      "eval_runtime": 13.2807,
      "eval_samples_per_second": 282.365,
      "eval_steps_per_second": 4.443,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 48.912818908691406,
      "learning_rate": 4.951556604879048e-06,
      "loss": 0.2632,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.86,
      "eval_loss": 0.4424697160720825,
      "eval_runtime": 13.1228,
      "eval_samples_per_second": 285.762,
      "eval_steps_per_second": 4.496,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 114.86393737792969,
      "learning_rate": 0.0,
      "loss": 0.2505,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8608,
      "eval_loss": 0.4482859671115875,
      "eval_runtime": 13.2334,
      "eval_samples_per_second": 283.374,
      "eval_steps_per_second": 4.458,
      "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
}