run-1/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.805045871559633,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-1/checkpoint-1054",
  "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": 6.5936455726623535,
      "learning_rate": 3.3598671878261075e-05,
      "loss": 2.2445,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.786697247706422,
      "eval_loss": 1.6360795497894287,
      "eval_runtime": 2.8227,
      "eval_samples_per_second": 308.927,
      "eval_steps_per_second": 2.48,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 2.5230881206493013e-05,
      "loss": 1.3366,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.805045871559633,
      "eval_loss": 1.3841164112091064,
      "eval_runtime": 2.8165,
      "eval_samples_per_second": 309.6,
      "eval_steps_per_second": 2.485,
      "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": 0.4702639641604981,
    "learning_rate": 4.199833984782634e-05,
    "num_train_epochs": 5,
    "temperature": 22
  }
}