trainer_state.json

{
  "best_metric": 0.21865835785865784,
  "best_model_checkpoint": "./vit-strawberries/checkpoint-80",
  "epoch": 16.0,
  "eval_steps": 40,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.0,
      "grad_norm": 0.7542967796325684,
      "learning_rate": 0.00018181818181818183,
      "loss": 0.4054,
      "step": 10
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.1642914116382599,
      "learning_rate": 0.00016363636363636366,
      "loss": 0.0477,
      "step": 20
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.09644770622253418,
      "learning_rate": 0.00014545454545454546,
      "loss": 0.0222,
      "step": 30
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.0752597227692604,
      "learning_rate": 0.00012727272727272728,
      "loss": 0.0107,
      "step": 40
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8846153846153846,
      "eval_loss": 0.38944709300994873,
      "eval_runtime": 33.1468,
      "eval_samples_per_second": 0.784,
      "eval_steps_per_second": 0.121,
      "step": 40
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.0672512948513031,
      "learning_rate": 0.00010909090909090909,
      "loss": 0.0085,
      "step": 50
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.055992938578128815,
      "learning_rate": 9.090909090909092e-05,
      "loss": 0.0073,
      "step": 60
    },
    {
      "epoch": 14.0,
      "grad_norm": 0.05227123200893402,
      "learning_rate": 7.272727272727273e-05,
      "loss": 0.0066,
      "step": 70
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.05070338025689125,
      "learning_rate": 5.4545454545454546e-05,
      "loss": 0.0061,
      "step": 80
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9230769230769231,
      "eval_loss": 0.21865835785865784,
      "eval_runtime": 16.5967,
      "eval_samples_per_second": 1.567,
      "eval_steps_per_second": 0.241,
      "step": 80
    }
  ],
  "logging_steps": 10,
  "max_steps": 110,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 22,
  "save_steps": 40,
  "total_flos": 9.175051570367693e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}