trainer_state.json

{
  "best_metric": 0.8858666666666667,
  "best_model_checkpoint": "./resnet_finetuned_models_dataset/CIFAR100/50_from_100/microsoft_resnet-101/model_idx_0015/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": 147.51644897460938,
      "learning_rate": 9.504844339512095e-05,
      "loss": 2.9348,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6893333333333334,
      "eval_loss": 1.201104998588562,
      "eval_runtime": 9.6827,
      "eval_samples_per_second": 387.288,
      "eval_steps_per_second": 6.093,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 54.793975830078125,
      "learning_rate": 8.117449009293668e-05,
      "loss": 0.8303,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8226666666666667,
      "eval_loss": 0.6036300659179688,
      "eval_runtime": 8.8863,
      "eval_samples_per_second": 421.996,
      "eval_steps_per_second": 6.639,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 71.83878326416016,
      "learning_rate": 6.112604669781572e-05,
      "loss": 0.5127,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8565333333333334,
      "eval_loss": 0.4763355553150177,
      "eval_runtime": 8.8647,
      "eval_samples_per_second": 423.028,
      "eval_steps_per_second": 6.656,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 50.52066421508789,
      "learning_rate": 3.887395330218429e-05,
      "loss": 0.3953,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8733333333333333,
      "eval_loss": 0.4242381751537323,
      "eval_runtime": 8.6957,
      "eval_samples_per_second": 431.249,
      "eval_steps_per_second": 6.785,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 51.31986618041992,
      "learning_rate": 1.8825509907063327e-05,
      "loss": 0.3197,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8792,
      "eval_loss": 0.39328446984291077,
      "eval_runtime": 8.7002,
      "eval_samples_per_second": 431.026,
      "eval_steps_per_second": 6.781,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 56.45174789428711,
      "learning_rate": 4.951556604879048e-06,
      "loss": 0.295,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8837333333333334,
      "eval_loss": 0.3996211588382721,
      "eval_runtime": 8.978,
      "eval_samples_per_second": 417.685,
      "eval_steps_per_second": 6.572,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 56.920982360839844,
      "learning_rate": 0.0,
      "loss": 0.2804,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8858666666666667,
      "eval_loss": 0.39302414655685425,
      "eval_runtime": 8.8475,
      "eval_samples_per_second": 423.85,
      "eval_steps_per_second": 6.669,
      "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
}