trainer_state.json

{
  "best_metric": 0.9405333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0524/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.4368906021118164,
      "learning_rate": 2.6666666666666667e-05,
      "loss": 1.188,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9165333333333333,
      "eval_loss": 0.342217355966568,
      "eval_runtime": 10.9795,
      "eval_samples_per_second": 341.547,
      "eval_steps_per_second": 5.374,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.40241539478302,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.1831,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9256,
      "eval_loss": 0.2581632435321808,
      "eval_runtime": 11.6915,
      "eval_samples_per_second": 320.746,
      "eval_steps_per_second": 5.046,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.09079994261264801,
      "learning_rate": 1.9999999999999998e-05,
      "loss": 0.0879,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9333333333333333,
      "eval_loss": 0.24043036997318268,
      "eval_runtime": 11.172,
      "eval_samples_per_second": 335.662,
      "eval_steps_per_second": 5.281,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.057769775390625,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0457,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9336,
      "eval_loss": 0.24038249254226685,
      "eval_runtime": 11.2682,
      "eval_samples_per_second": 332.794,
      "eval_steps_per_second": 5.236,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.11284955590963364,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0249,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.936,
      "eval_loss": 0.2474948912858963,
      "eval_runtime": 11.3736,
      "eval_samples_per_second": 329.712,
      "eval_steps_per_second": 5.187,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.1586252748966217,
      "learning_rate": 9.999999999999999e-06,
      "loss": 0.0133,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9370666666666667,
      "eval_loss": 0.25553789734840393,
      "eval_runtime": 11.4947,
      "eval_samples_per_second": 326.236,
      "eval_steps_per_second": 5.133,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 6.676676676676677e-06,
      "loss": 0.008,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9368,
      "eval_loss": 0.2485409826040268,
      "eval_runtime": 10.9948,
      "eval_samples_per_second": 341.07,
      "eval_steps_per_second": 5.366,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.12917065620422363,
      "learning_rate": 3.3433433433433436e-06,
      "loss": 0.0053,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9392,
      "eval_loss": 0.23959769308567047,
      "eval_runtime": 10.9903,
      "eval_samples_per_second": 341.211,
      "eval_steps_per_second": 5.368,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.6660424470901489,
      "learning_rate": 1.001001001001001e-08,
      "loss": 0.0043,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9405333333333333,
      "eval_loss": 0.25492385029792786,
      "eval_runtime": 11.04,
      "eval_samples_per_second": 339.673,
      "eval_steps_per_second": 5.344,
      "step": 2997
    }
  ],
  "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": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}