run-10/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": 0.7502514123916626,
      "eval_loss": 0.47628718614578247,
      "eval_runtime": 9.4864,
      "eval_samples_per_second": 314.449,
      "eval_steps_per_second": 19.712,
      "step": 373
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 11.133524894714355,
      "learning_rate": 1.3824026202690149e-05,
      "loss": 0.5239,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7720415592193604,
      "eval_loss": 0.4691523015499115,
      "eval_runtime": 9.6019,
      "eval_samples_per_second": 310.669,
      "eval_steps_per_second": 19.475,
      "step": 746
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 12.096076011657715,
      "learning_rate": 1.1684084066050808e-05,
      "loss": 0.2819,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7730472683906555,
      "eval_loss": 0.6077698469161987,
      "eval_runtime": 9.4502,
      "eval_samples_per_second": 315.654,
      "eval_steps_per_second": 19.788,
      "step": 1119
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7696949243545532,
      "eval_loss": 0.8962089419364929,
      "eval_runtime": 9.4938,
      "eval_samples_per_second": 314.204,
      "eval_steps_per_second": 19.697,
      "step": 1492
    },
    {
      "epoch": 4.021447721179625,
      "grad_norm": 1.9103776216506958,
      "learning_rate": 9.544141929411465e-06,
      "loss": 0.1205,
      "step": 1500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7790814638137817,
      "eval_loss": 1.138683795928955,
      "eval_runtime": 9.5013,
      "eval_samples_per_second": 313.959,
      "eval_steps_per_second": 19.682,
      "step": 1865
    },
    {
      "epoch": 5.361930294906166,
      "grad_norm": 12.819802284240723,
      "learning_rate": 7.4041997927721225e-06,
      "loss": 0.0447,
      "step": 2000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7750586867332458,
      "eval_loss": 1.263724446296692,
      "eval_runtime": 9.4719,
      "eval_samples_per_second": 314.933,
      "eval_steps_per_second": 19.743,
      "step": 2238
    },
    {
      "epoch": 6.702412868632708,
      "grad_norm": 0.741857647895813,
      "learning_rate": 5.264257656132781e-06,
      "loss": 0.0301,
      "step": 2500
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7790814638137817,
      "eval_loss": 1.3301361799240112,
      "eval_runtime": 9.4495,
      "eval_samples_per_second": 315.677,
      "eval_steps_per_second": 19.789,
      "step": 2611
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.7787462472915649,
      "eval_loss": 1.3093613386154175,
      "eval_runtime": 9.6096,
      "eval_samples_per_second": 310.419,
      "eval_steps_per_second": 19.46,
      "step": 2984
    },
    {
      "epoch": 8.04289544235925,
      "grad_norm": 0.6469907760620117,
      "learning_rate": 3.124315519493439e-06,
      "loss": 0.0153,
      "step": 3000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.7770700454711914,
      "eval_loss": 1.3546313047409058,
      "eval_runtime": 9.6011,
      "eval_samples_per_second": 310.692,
      "eval_steps_per_second": 19.477,
      "step": 3357
    },
    {
      "epoch": 9.383378016085791,
      "grad_norm": 4.288186550140381,
      "learning_rate": 9.843733828540973e-07,
      "loss": 0.0116,
      "step": 3500
    }
  ],
  "logging_steps": 500,
  "max_steps": 3730,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2.000097781399735e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.596396833932949e-05,
    "per_device_train_batch_size": 32
  }
}