run-4/checkpoint-801/trainer_state.json

{
  "best_metric": 0.5776590704917908,
  "best_model_checkpoint": "BERT-WMM/run-4/checkpoint-534",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 801,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7577464788732394,
      "eval_loss": 0.6081846952438354,
      "eval_runtime": 2.4392,
      "eval_samples_per_second": 873.231,
      "eval_steps_per_second": 54.936,
      "step": 267
    },
    {
      "epoch": 1.87,
      "grad_norm": 9.532360076904297,
      "learning_rate": 9.65168237740807e-06,
      "loss": 0.6057,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7657276995305164,
      "eval_loss": 0.5776590704917908,
      "eval_runtime": 2.4594,
      "eval_samples_per_second": 866.069,
      "eval_steps_per_second": 54.485,
      "step": 534
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7690140845070422,
      "eval_loss": 0.6647635102272034,
      "eval_runtime": 2.4523,
      "eval_samples_per_second": 868.556,
      "eval_steps_per_second": 54.642,
      "step": 801
    }
  ],
  "logging_steps": 500,
  "max_steps": 1335,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 349221394826640.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "num_train_epochs": 5
  }
}