Apply yapf formatting

Author: marpng

joeynmt/prediction.py

@@ -50,8 +50,14 @@ def predict(
     num_workers: int = 0,
     cfg: Dict = None,
     fp16: bool = False,
-) -> Tuple[Dict[str, float], List[str], List[str], List[List[str]], List[np.ndarray],
-           List[np.ndarray], ]:
+) -> Tuple[
+        Dict[str, float],
+        List[str],
+        List[str],
+        List[List[str]],
+        List[np.ndarray],
+        List[np.ndarray],
+]:
     """
     Generates translations for the given data.
     If `compute_loss` is True and references are given, also computes the loss.

joeynmt/tokenizers.py

@@ -309,14 +309,14 @@ def set_vocab(self, itos: List[str]) -> None:
     def copy_cfg_file(self, model_dir: Path) -> None:
         """this does nothing at the moment, uncommenting the original leads to training error"""
         useless = "do nothing"
+
     #def copy_cfg_file(self, model_dir: Path) -> None:
     #    """Copy config file to model_dir"""
     #    print(f"DEBUG: \n self.codes: \t{self.codes}\n \
     #          self.codes.name:\t{self.codes.name}\nmodel_dir:\t{model_dir}\n \
     #          the other thing:\t{(model_dir / self.codes.name).as_posix()}")
     #    shutil.copy2(self.codes, (model_dir / self.codes.name).as_posix())
 
-
     def __repr__(self):
         return (f"{self.__class__.__name__}(level={self.level}, "
                 f"lowercase={self.lowercase}, normalize={self.normalize}, "

joeynmt/training.py

@@ -578,9 +578,9 @@ def _train_step(self, batch: Batch) -> Tensor:
         #with torch.autocast(device_type=self.device.type,
         #                    dtype=self.dtype,
         #                    enabled=self.fp16):
-            # get loss (run as during training with teacher forcing)
-            #batch_loss, _, _, correct_tokens = self.model(return_type="loss",
-            #                                                 **vars(batch))
+        # get loss (run as during training with teacher forcing)
+        #batch_loss, _, _, correct_tokens = self.model(return_type="loss",
+        #                                                 **vars(batch))
         batch_loss, _, _, correct_tokens = self.model(return_type="loss",
                                                           **vars(batch))