run-4/checkpoint-1263/trainer_state.json

{
  "best_metric": 0.911697247706422,
  "best_model_checkpoint": "bert-base-uncased-finetuned-sst2/run-4/checkpoint-421",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1263,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.911697247706422,
      "eval_loss": 0.22564278542995453,
      "eval_runtime": 2.3507,
      "eval_samples_per_second": 370.952,
      "eval_steps_per_second": 23.397,
      "step": 421
    },
    {
      "epoch": 1.19,
      "grad_norm": 22.03980827331543,
      "learning_rate": 1.3592312257395198e-05,
      "loss": 0.3224,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.908256880733945,
      "eval_loss": 0.28558045625686646,
      "eval_runtime": 2.3703,
      "eval_samples_per_second": 367.886,
      "eval_steps_per_second": 23.204,
      "step": 842
    },
    {
      "epoch": 2.38,
      "grad_norm": 6.299999713897705,
      "learning_rate": 7.85231552707628e-06,
      "loss": 0.1554,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9071100917431193,
      "eval_loss": 0.4138067066669464,
      "eval_runtime": 2.3765,
      "eval_samples_per_second": 366.925,
      "eval_steps_per_second": 23.143,
      "step": 1263
    }
  ],
  "logging_steps": 500,
  "max_steps": 1684,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 290833299485880.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.9332308987714115e-05,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 16,
    "seed": 7
  }
}