run-8/checkpoint-1581/trainer_state.json

{
  "best_metric": 0.9036697247706422,
  "best_model_checkpoint": "electra-distilled-sst\\run-8\\checkpoint-527",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1581,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 25.336427688598633,
      "learning_rate": 8.271664431386432e-05,
      "loss": 0.94,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9036697247706422,
      "eval_loss": 0.6703569889068604,
      "eval_runtime": 1.4486,
      "eval_samples_per_second": 601.968,
      "eval_steps_per_second": 4.832,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 11.467299461364746,
      "learning_rate": 7.352784339619853e-05,
      "loss": 0.4737,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8990825688073395,
      "eval_loss": 0.8164154887199402,
      "eval_runtime": 1.4496,
      "eval_samples_per_second": 601.528,
      "eval_steps_per_second": 4.829,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": 44.32056427001953,
      "learning_rate": 6.433904247853274e-05,
      "loss": 0.3435,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8979357798165137,
      "eval_loss": 0.8911434412002563,
      "eval_runtime": 1.4479,
      "eval_samples_per_second": 602.254,
      "eval_steps_per_second": 4.835,
      "step": 1581
    }
  ],
  "logging_steps": 500,
  "max_steps": 5270,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 561038813831964.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7393993567935025,
    "learning_rate": 9.188800917665787e-05,
    "num_train_epochs": 10,
    "temperature": 6
  }
}