{
  "best_metric": 0.8302752293577982,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-12/checkpoint-527",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 3689,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 9.690923690795898,
      "learning_rate": 0.0003321213172401304,
      "loss": 1.3202,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8302752293577982,
      "eval_loss": 1.4912840127944946,
      "eval_runtime": 2.8077,
      "eval_samples_per_second": 310.575,
      "eval_steps_per_second": 2.493,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 8.249791145324707,
      "learning_rate": 0.0002767677643667753,
      "loss": 0.5967,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8107798165137615,
      "eval_loss": 1.6795676946640015,
      "eval_runtime": 2.8151,
      "eval_samples_per_second": 309.763,
      "eval_steps_per_second": 2.487,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": NaN,
      "learning_rate": 0.00022151924669811353,
      "loss": 0.4141,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8107798165137615,
      "eval_loss": 1.6723682880401611,
      "eval_runtime": 2.8228,
      "eval_samples_per_second": 308.915,
      "eval_steps_per_second": 2.48,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.678206443786621,
      "learning_rate": 0.00016616569382475846,
      "loss": 0.3104,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.823394495412844,
      "eval_loss": 1.6141194105148315,
      "eval_runtime": 2.8102,
      "eval_samples_per_second": 310.299,
      "eval_steps_per_second": 2.491,
      "step": 2108
    },
    {
      "epoch": 5.0,
      "grad_norm": 12.198997497558594,
      "learning_rate": 0.0001108121409514034,
      "loss": 0.2449,
      "step": 2635
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8165137614678899,
      "eval_loss": 1.7016170024871826,
      "eval_runtime": 2.8147,
      "eval_samples_per_second": 309.804,
      "eval_steps_per_second": 2.487,
      "step": 2635
    },
    {
      "epoch": 6.0,
      "grad_norm": 13.61044692993164,
      "learning_rate": 5.545858807804834e-05,
      "loss": 0.2063,
      "step": 3162
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8188073394495413,
      "eval_loss": 1.7187268733978271,
      "eval_runtime": 2.8199,
      "eval_samples_per_second": 309.227,
      "eval_steps_per_second": 2.482,
      "step": 3162
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.553529739379883,
      "learning_rate": 1.0503520469327336e-07,
      "loss": 0.1803,
      "step": 3689
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8222477064220184,
      "eval_loss": 1.734584093093872,
      "eval_runtime": 2.8212,
      "eval_samples_per_second": 309.087,
      "eval_steps_per_second": 2.481,
      "step": 3689
    }
  ],
  "logging_steps": 500,
  "max_steps": 3689,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 56628612372420.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.3539642822905392,
    "learning_rate": 0.00038747487011348545,
    "num_train_epochs": 7,
    "temperature": 26
  }
}