trainer_state.json

{
  "best_metric": 0.07614374160766602,
  "best_model_checkpoint": "./vit-base-16-thesis-demo-PH2/checkpoint-50",
  "epoch": 4.0,
  "eval_steps": 50,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "learning_rate": 0.000175,
      "loss": 0.3313,
      "step": 10
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.2035,
      "step": 20
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.000125,
      "loss": 0.1033,
      "step": 30
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0001,
      "loss": 0.0258,
      "step": 40
    },
    {
      "epoch": 2.5,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.044,
      "step": 50
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.984375,
      "eval_f1": 0.984375,
      "eval_loss": 0.07614374160766602,
      "eval_precision": 0.984375,
      "eval_recall": 0.984375,
      "eval_runtime": 1.566,
      "eval_samples_per_second": 40.868,
      "eval_steps_per_second": 5.109,
      "step": 50
    },
    {
      "epoch": 3.0,
      "learning_rate": 5e-05,
      "loss": 0.0182,
      "step": 60
    },
    {
      "epoch": 3.5,
      "learning_rate": 2.5e-05,
      "loss": 0.0309,
      "step": 70
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0,
      "loss": 0.0116,
      "step": 80
    },
    {
      "epoch": 4.0,
      "step": 80,
      "total_flos": 9.918974670667776e+16,
      "train_loss": 0.09609116911888123,
      "train_runtime": 67.2316,
      "train_samples_per_second": 19.039,
      "train_steps_per_second": 1.19
    }
  ],
  "logging_steps": 10,
  "max_steps": 80,
  "num_train_epochs": 4,
  "save_steps": 50,
  "total_flos": 9.918974670667776e+16,
  "trial_name": null,
  "trial_params": null
}