run-3/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": 0.6443178057670593,
      "eval_loss": 0.6200888156890869,
      "eval_runtime": 9.4558,
      "eval_samples_per_second": 315.469,
      "eval_steps_per_second": 19.776,
      "step": 373
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 11.498894691467285,
      "learning_rate": 1.0365748013479516e-05,
      "loss": 0.6738,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7539389729499817,
      "eval_loss": 0.47277966141700745,
      "eval_runtime": 9.3806,
      "eval_samples_per_second": 317.998,
      "eval_steps_per_second": 19.935,
      "step": 746
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 9.671843528747559,
      "learning_rate": 8.761143057832532e-06,
      "loss": 0.4763,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7717063426971436,
      "eval_loss": 0.5170367360115051,
      "eval_runtime": 9.2678,
      "eval_samples_per_second": 321.867,
      "eval_steps_per_second": 20.177,
      "step": 1119
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7787462472915649,
      "eval_loss": 0.5018635392189026,
      "eval_runtime": 9.2787,
      "eval_samples_per_second": 321.49,
      "eval_steps_per_second": 20.154,
      "step": 1492
    },
    {
      "epoch": 4.021447721179625,
      "grad_norm": 22.328227996826172,
      "learning_rate": 7.156538102185547e-06,
      "loss": 0.2904,
      "step": 1500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7800871729850769,
      "eval_loss": 0.6816526651382446,
      "eval_runtime": 9.2909,
      "eval_samples_per_second": 321.068,
      "eval_steps_per_second": 20.127,
      "step": 1865
    },
    {
      "epoch": 5.361930294906166,
      "grad_norm": 8.851292610168457,
      "learning_rate": 5.551933146538564e-06,
      "loss": 0.1552,
      "step": 2000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.782433807849884,
      "eval_loss": 0.8527898192405701,
      "eval_runtime": 9.405,
      "eval_samples_per_second": 317.172,
      "eval_steps_per_second": 19.883,
      "step": 2238
    },
    {
      "epoch": 6.702412868632708,
      "grad_norm": 20.3522891998291,
      "learning_rate": 3.94732819089158e-06,
      "loss": 0.1004,
      "step": 2500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7747234106063843,
      "eval_loss": 0.9924494624137878,
      "eval_runtime": 9.2681,
      "eval_samples_per_second": 321.856,
      "eval_steps_per_second": 20.177,
      "step": 2611
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7780757546424866,
      "eval_loss": 1.0592091083526611,
      "eval_runtime": 9.4254,
      "eval_samples_per_second": 316.486,
      "eval_steps_per_second": 19.84,
      "step": 2984
    },
    {
      "epoch": 8.04289544235925,
      "grad_norm": 6.993451118469238,
      "learning_rate": 2.342723235244596e-06,
      "loss": 0.0685,
      "step": 3000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7794166803359985,
      "eval_loss": 1.097713589668274,
      "eval_runtime": 9.4231,
      "eval_samples_per_second": 316.561,
      "eval_steps_per_second": 19.845,
      "step": 3357
    },
    {
      "epoch": 9.383378016085791,
      "grad_norm": 36.58711242675781,
      "learning_rate": 7.381182795976126e-07,
      "loss": 0.0513,
      "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": 1.1970352969126499e-05,
    "per_device_train_batch_size": 32
  }
}