{
  "best_metric": 0.575,
  "best_model_checkpoint": "NEW-finetuned/checkpoint-32",
  "epoch": 3.9298245614035086,
  "eval_steps": 500,
  "global_step": 32,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.98,
      "eval_accuracy": 0.295,
      "eval_loss": 1.5839855670928955,
      "eval_runtime": 5.3781,
      "eval_samples_per_second": 37.188,
      "eval_steps_per_second": 1.302,
      "step": 8
    },
    {
      "epoch": 1.23,
      "grad_norm": 3.0023796558380127,
      "learning_rate": 7.857142857142858e-06,
      "loss": 1.606,
      "step": 10
    },
    {
      "epoch": 1.96,
      "eval_accuracy": 0.495,
      "eval_loss": 1.5293787717819214,
      "eval_runtime": 3.4271,
      "eval_samples_per_second": 58.359,
      "eval_steps_per_second": 2.043,
      "step": 16
    },
    {
      "epoch": 2.46,
      "grad_norm": 3.5246360301971436,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 1.5151,
      "step": 20
    },
    {
      "epoch": 2.95,
      "eval_accuracy": 0.55,
      "eval_loss": 1.4795724153518677,
      "eval_runtime": 3.4766,
      "eval_samples_per_second": 57.528,
      "eval_steps_per_second": 2.013,
      "step": 24
    },
    {
      "epoch": 3.68,
      "grad_norm": 3.501955986022949,
      "learning_rate": 7.142857142857143e-07,
      "loss": 1.4497,
      "step": 30
    },
    {
      "epoch": 3.93,
      "eval_accuracy": 0.575,
      "eval_loss": 1.4569343328475952,
      "eval_runtime": 3.533,
      "eval_samples_per_second": 56.609,
      "eval_steps_per_second": 1.981,
      "step": 32
    },
    {
      "epoch": 3.93,
      "step": 32,
      "total_flos": 1.7639288313008947e+17,
      "train_loss": 1.5180421471595764,
      "train_runtime": 155.1895,
      "train_samples_per_second": 46.395,
      "train_steps_per_second": 0.206
    }
  ],
  "logging_steps": 10,
  "max_steps": 32,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 1.7639288313008947e+17,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}