trainer_state.json

{
  "best_metric": 0.5133863687515259,
  "best_model_checkpoint": "./vit-manuscripts/checkpoint-34",
  "epoch": 1.0,
  "global_step": 34,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.29,
      "learning_rate": 7.05720474130633e-05,
      "loss": 1.8758,
      "step": 10
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.117564276235853e-05,
      "loss": 0.6165,
      "step": 20
    },
    {
      "epoch": 0.88,
      "learning_rate": 8.512107998897363e-06,
      "loss": 0.5303,
      "step": 30
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5133863687515259,
      "eval_runtime": 844.4272,
      "eval_samples_per_second": 0.888,
      "eval_steps_per_second": 0.007,
      "step": 34
    },
    {
      "epoch": 1.0,
      "step": 34,
      "total_flos": 4.2955421515776e+17,
      "train_loss": 0.9502218470853918,
      "train_runtime": 5818.3708,
      "train_samples_per_second": 0.73,
      "train_steps_per_second": 0.006
    }
  ],
  "max_steps": 34,
  "num_train_epochs": 1,
  "total_flos": 4.2955421515776e+17,
  "trial_name": null,
  "trial_params": null
}