trainer_state.json

{
  "best_metric": 0.0366968959569931,
  "best_model_checkpoint": "vit-large-patch16-224-dungeon-geo-morphs-007/checkpoint-10",
  "epoch": 20.571428571428573,
  "eval_steps": 10,
  "global_step": 31,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 6.571428571428571,
      "grad_norm": 0.000408386840717867,
      "learning_rate": 1.555555555555556e-05,
      "loss": 0.0,
      "step": 10
    },
    {
      "epoch": 6.571428571428571,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 0.0366968959569931,
      "eval_runtime": 0.6753,
      "eval_samples_per_second": 53.307,
      "eval_steps_per_second": 7.404,
      "step": 10
    },
    {
      "epoch": 13.285714285714286,
      "grad_norm": 8.39885396999307e-05,
      "learning_rate": 8.148148148148148e-06,
      "loss": 0.0,
      "step": 20
    },
    {
      "epoch": 13.285714285714286,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 0.06323710083961487,
      "eval_runtime": 0.4956,
      "eval_samples_per_second": 72.642,
      "eval_steps_per_second": 10.089,
      "step": 20
    },
    {
      "epoch": 19.857142857142858,
      "grad_norm": 6.093415504437871e-05,
      "learning_rate": 7.407407407407407e-07,
      "loss": 0.0,
      "step": 30
    },
    {
      "epoch": 19.857142857142858,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 0.0674421563744545,
      "eval_runtime": 0.5889,
      "eval_samples_per_second": 61.134,
      "eval_steps_per_second": 8.491,
      "step": 30
    },
    {
      "epoch": 20.571428571428573,
      "step": 31,
      "total_flos": 2.717436247668818e+17,
      "train_loss": 4.973743231065812e-06,
      "train_runtime": 517.3884,
      "train_samples_per_second": 2.996,
      "train_steps_per_second": 0.06
    }
  ],
  "logging_steps": 10,
  "max_steps": 31,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 31,
  "save_steps": 10,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.01
      },
      "attributes": {
        "early_stopping_patience_counter": 2
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.717436247668818e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}