{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.684563758389262,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.34,
      "grad_norm": 2.6059138774871826,
      "learning_rate": 4.440715883668904e-05,
      "loss": 0.6515,
      "step": 500
    },
    {
      "epoch": 0.67,
      "grad_norm": 3.4959137439727783,
      "learning_rate": 3.881431767337807e-05,
      "loss": 0.6411,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7209302325581395,
      "eval_f1": 0.0,
      "eval_loss": 0.6031970381736755,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 1.3563,
      "eval_samples_per_second": 634.099,
      "eval_steps_per_second": 79.631,
      "step": 1490
    },
    {
      "epoch": 1.01,
      "grad_norm": 0.5044479370117188,
      "learning_rate": 3.3221476510067115e-05,
      "loss": 0.6488,
      "step": 1500
    },
    {
      "epoch": 1.34,
      "grad_norm": 1.092705249786377,
      "learning_rate": 2.7628635346756153e-05,
      "loss": 0.6487,
      "step": 2000
    },
    {
      "epoch": 1.68,
      "grad_norm": 1.8183097839355469,
      "learning_rate": 2.2035794183445192e-05,
      "loss": 0.6313,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7209302325581395,
      "eval_f1": 0.0,
      "eval_loss": 0.6042766571044922,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 1.5777,
      "eval_samples_per_second": 545.08,
      "eval_steps_per_second": 68.452,
      "step": 2980
    },
    {
      "epoch": 2.01,
      "grad_norm": 1.3732738494873047,
      "learning_rate": 1.644295302013423e-05,
      "loss": 0.6426,
      "step": 3000
    },
    {
      "epoch": 2.35,
      "grad_norm": 1.1144658327102661,
      "learning_rate": 1.0850111856823267e-05,
      "loss": 0.6399,
      "step": 3500
    },
    {
      "epoch": 2.68,
      "grad_norm": 0.6613803505897522,
      "learning_rate": 5.257270693512305e-06,
      "loss": 0.6454,
      "step": 4000
    }
  ],
  "logging_steps": 500,
  "max_steps": 4470,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1032460114563360.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}