run-28/checkpoint-594/trainer_state.json

{
  "best_metric": 0.6514851485148515,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-28/checkpoint-594",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 594,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.3590941429138184,
      "learning_rate": 0.00033268893750406807,
      "loss": 0.6415,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5861386138613861,
      "eval_f1": 0.473551637279597,
      "eval_loss": 0.6401379108428955,
      "eval_mcc": 0.1894629867942258,
      "eval_precision": 0.6482758620689655,
      "eval_recall": 0.373015873015873,
      "eval_runtime": 0.9371,
      "eval_samples_per_second": 538.913,
      "eval_steps_per_second": 17.074,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.476292848587036,
      "learning_rate": 0.00028516194643205835,
      "loss": 0.613,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6514851485148515,
      "eval_f1": 0.68,
      "eval_loss": 0.5918833017349243,
      "eval_mcc": 0.30837572082433856,
      "eval_precision": 0.62751677852349,
      "eval_recall": 0.7420634920634921,
      "eval_runtime": 0.9445,
      "eval_samples_per_second": 534.7,
      "eval_steps_per_second": 16.941,
      "step": 594
    }
  ],
  "logging_steps": 500,
  "max_steps": 2376,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 1461402531960.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9092941270097592,
    "learning_rate": 0.0003802159285760778,
    "num_train_epochs": 8,
    "temperature": 11
  }
}