run-1/checkpoint-1684/trainer_state.json

{
  "best_metric": 0.9139908256880734,
  "best_model_checkpoint": "bert-base-uncased-finetuned-sst2/run-1/checkpoint-1684",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1684,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.59,
      "grad_norm": 5.992064476013184,
      "learning_rate": 2.846915930651389e-05,
      "loss": 0.4119,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8646788990825688,
      "eval_loss": 0.40675637125968933,
      "eval_runtime": 2.35,
      "eval_samples_per_second": 371.063,
      "eval_steps_per_second": 23.404,
      "step": 842
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.04706716537475586,
      "learning_rate": 1.6446710275046875e-05,
      "loss": 0.2691,
      "step": 1000
    },
    {
      "epoch": 1.78,
      "grad_norm": 0.055366575717926025,
      "learning_rate": 4.424261243579861e-06,
      "loss": 0.1703,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9139908256880734,
      "eval_loss": 0.3520638048648834,
      "eval_runtime": 2.3883,
      "eval_samples_per_second": 365.121,
      "eval_steps_per_second": 23.029,
      "step": 1684
    }
  ],
  "logging_steps": 500,
  "max_steps": 1684,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 185893613388420.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 4.0491608337980904e-05,
    "num_train_epochs": 2,
    "per_device_train_batch_size": 8,
    "seed": 26
  }
}