run-24/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.8337155963302753,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-24/checkpoint-527",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 1054,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 17.628143310546875,
      "learning_rate": 0.00018567149458255904,
      "loss": 1.3968,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8337155963302753,
      "eval_loss": 1.3282397985458374,
      "eval_runtime": 2.8214,
      "eval_samples_per_second": 309.061,
      "eval_steps_per_second": 2.481,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 10.100229263305664,
      "learning_rate": 0.00013925362093691925,
      "loss": 0.6522,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8291284403669725,
      "eval_loss": 1.401127576828003,
      "eval_runtime": 2.8077,
      "eval_samples_per_second": 310.572,
      "eval_steps_per_second": 2.493,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 2635,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 16141525436100.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.4205009910693592,
    "learning_rate": 0.00023208936822819877,
    "num_train_epochs": 5,
    "temperature": 21
  }
}