run-0/checkpoint-1054/trainer_state.json

{
  "best_metric": 0.8348623853211009,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-0/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,
      "learning_rate": 0.00017521452501048488,
      "loss": 1.4337,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8211009174311926,
      "eval_loss": 1.309804081916809,
      "eval_runtime": 7.7796,
      "eval_samples_per_second": 112.088,
      "eval_steps_per_second": 0.9,
      "step": 527
    },
    {
      "epoch": 2.0,
      "learning_rate": 8.760726250524244e-05,
      "loss": 0.6599,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8348623853211009,
      "eval_loss": 1.4746311902999878,
      "eval_runtime": 10.2236,
      "eval_samples_per_second": 85.293,
      "eval_steps_per_second": 0.685,
      "step": 1054
    }
  ],
  "logging_steps": 500,
  "max_steps": 1581,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 16162868129520.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.3828930992000458,
    "learning_rate": 0.0002628217875157273,
    "num_train_epochs": 3,
    "temperature": 27
  }
}