run-4/checkpoint-4000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.361930294906166,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.6702412868632708,
      "grad_norm": 14.762006759643555,
      "learning_rate": 6.627475083076973e-06,
      "loss": 0.6513,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6557157039642334,
      "eval_loss": 0.589284360408783,
      "eval_runtime": 17.191,
      "eval_samples_per_second": 173.521,
      "eval_steps_per_second": 10.878,
      "step": 746
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 12.195761680603027,
      "learning_rate": 6.151363367338686e-06,
      "loss": 0.5889,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6714716553688049,
      "eval_loss": 0.5709448456764221,
      "eval_runtime": 13.4458,
      "eval_samples_per_second": 221.853,
      "eval_steps_per_second": 13.908,
      "step": 1492
    },
    {
      "epoch": 2.0107238605898123,
      "grad_norm": 8.70524787902832,
      "learning_rate": 5.675251651600397e-06,
      "loss": 0.5639,
      "step": 1500
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 19.87879753112793,
      "learning_rate": 5.199139935862109e-06,
      "loss": 0.5108,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6741535663604736,
      "eval_loss": 0.5880420207977295,
      "eval_runtime": 17.878,
      "eval_samples_per_second": 166.853,
      "eval_steps_per_second": 10.46,
      "step": 2238
    },
    {
      "epoch": 3.351206434316354,
      "grad_norm": 16.54958152770996,
      "learning_rate": 4.723028220123821e-06,
      "loss": 0.484,
      "step": 2500
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6832048296928406,
      "eval_loss": 0.5972875356674194,
      "eval_runtime": 13.3249,
      "eval_samples_per_second": 223.866,
      "eval_steps_per_second": 14.034,
      "step": 2984
    },
    {
      "epoch": 4.021447721179625,
      "grad_norm": 18.406064987182617,
      "learning_rate": 4.246916504385532e-06,
      "loss": 0.4636,
      "step": 3000
    },
    {
      "epoch": 4.6916890080428955,
      "grad_norm": 21.423734664916992,
      "learning_rate": 3.770804788647244e-06,
      "loss": 0.4134,
      "step": 3500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6915856599807739,
      "eval_loss": 0.6475616693496704,
      "eval_runtime": 18.486,
      "eval_samples_per_second": 161.366,
      "eval_steps_per_second": 10.116,
      "step": 3730
    },
    {
      "epoch": 5.361930294906166,
      "grad_norm": 36.32148361206055,
      "learning_rate": 3.294693072908955e-06,
      "loss": 0.3833,
      "step": 4000
    }
  ],
  "logging_steps": 500,
  "max_steps": 7460,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2722096843486932.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 7.103586798815262e-06,
    "per_device_train_batch_size": 16
  }
}