run-2/checkpoint-1340/trainer_state.json

{
  "best_metric": 0.4286063983911785,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-2/checkpoint-1340",
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 1340,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.5596779584884644,
      "eval_matthews_correlation": 0.06558874629318973,
      "eval_runtime": 0.7398,
      "eval_samples_per_second": 1409.869,
      "eval_steps_per_second": 89.215,
      "step": 268
    },
    {
      "epoch": 1.87,
      "grad_norm": 3.7759640216827393,
      "learning_rate": 2.1988379475637637e-06,
      "loss": 0.5503,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.5253149271011353,
      "eval_matthews_correlation": 0.3853198145814999,
      "eval_runtime": 0.9856,
      "eval_samples_per_second": 1058.272,
      "eval_steps_per_second": 66.966,
      "step": 536
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5222654938697815,
      "eval_matthews_correlation": 0.4156507632951837,
      "eval_runtime": 0.7857,
      "eval_samples_per_second": 1327.408,
      "eval_steps_per_second": 83.997,
      "step": 804
    },
    {
      "epoch": 3.73,
      "grad_norm": 5.163796424865723,
      "learning_rate": 8.900058359186663e-07,
      "loss": 0.4554,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.5052011609077454,
      "eval_matthews_correlation": 0.42108781728525657,
      "eval_runtime": 1.6038,
      "eval_samples_per_second": 650.334,
      "eval_steps_per_second": 41.153,
      "step": 1072
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.5154977440834045,
      "eval_matthews_correlation": 0.4286063983911785,
      "eval_runtime": 0.848,
      "eval_samples_per_second": 1229.909,
      "eval_steps_per_second": 77.827,
      "step": 1340
    }
  ],
  "logging_steps": 500,
  "max_steps": 1340,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 191095347752964.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 3.5076700592088608e-06,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 32,
    "seed": 32
  }
}