run-4/checkpoint-3500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.383378016085791,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7321488261222839,
      "eval_loss": 0.5215846300125122,
      "eval_runtime": 9.3633,
      "eval_samples_per_second": 318.585,
      "eval_steps_per_second": 19.972,
      "step": 373
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 13.217938423156738,
      "learning_rate": 2.418286940875578e-06,
      "loss": 0.5967,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7723767757415771,
      "eval_loss": 0.45723703503608704,
      "eval_runtime": 9.2866,
      "eval_samples_per_second": 321.216,
      "eval_steps_per_second": 20.137,
      "step": 746
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 15.163392066955566,
      "learning_rate": 2.043939117210628e-06,
      "loss": 0.4683,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.796513557434082,
      "eval_loss": 0.4389999508857727,
      "eval_runtime": 9.2689,
      "eval_samples_per_second": 321.829,
      "eval_steps_per_second": 20.175,
      "step": 1119
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8055648803710938,
      "eval_loss": 0.45021912455558777,
      "eval_runtime": 9.4524,
      "eval_samples_per_second": 315.58,
      "eval_steps_per_second": 19.783,
      "step": 1492
    },
    {
      "epoch": 4.021447721179625,
      "grad_norm": 33.55284118652344,
      "learning_rate": 1.6695912935456774e-06,
      "loss": 0.3783,
      "step": 1500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8079115152359009,
      "eval_loss": 0.4486864507198334,
      "eval_runtime": 9.329,
      "eval_samples_per_second": 319.756,
      "eval_steps_per_second": 20.045,
      "step": 1865
    },
    {
      "epoch": 5.361930294906166,
      "grad_norm": 17.95193862915039,
      "learning_rate": 1.2952434698807272e-06,
      "loss": 0.3049,
      "step": 2000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8139457106590271,
      "eval_loss": 0.44699645042419434,
      "eval_runtime": 9.3232,
      "eval_samples_per_second": 319.955,
      "eval_steps_per_second": 20.058,
      "step": 2238
    },
    {
      "epoch": 6.702412868632708,
      "grad_norm": 19.94598388671875,
      "learning_rate": 9.208956462157772e-07,
      "loss": 0.2663,
      "step": 2500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8162923455238342,
      "eval_loss": 0.4586087167263031,
      "eval_runtime": 9.3905,
      "eval_samples_per_second": 317.661,
      "eval_steps_per_second": 19.914,
      "step": 2611
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8169627785682678,
      "eval_loss": 0.46715453267097473,
      "eval_runtime": 9.458,
      "eval_samples_per_second": 315.395,
      "eval_steps_per_second": 19.772,
      "step": 2984
    },
    {
      "epoch": 8.04289544235925,
      "grad_norm": 21.875141143798828,
      "learning_rate": 5.465478225508271e-07,
      "loss": 0.2394,
      "step": 3000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8183037042617798,
      "eval_loss": 0.4889311194419861,
      "eval_runtime": 9.4511,
      "eval_samples_per_second": 315.626,
      "eval_steps_per_second": 19.786,
      "step": 3357
    },
    {
      "epoch": 9.383378016085791,
      "grad_norm": 31.631711959838867,
      "learning_rate": 1.7219999888587704e-07,
      "loss": 0.2114,
      "step": 3500
    }
  ],
  "logging_steps": 500,
  "max_steps": 3730,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.9626746750006364e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 2.7926347645405278e-06,
    "per_device_train_batch_size": 32
  }
}