Replace assert statements with exceptions (#13871) (#13901)

* Replace assert statements with exceptions (#13871)

* Change f-strings when not needed (flake8)

* Replace assert statements with exceptions (#13871)

* Change f-strings when not needed (flake8)

* Improve error message as suggested by reviewer

* Fix identation bug

* Fix style errors

src/transformers/generation_beam_search.py

@@ -214,7 +214,17 @@ class BeamSearchScorer(BeamScorer):
     ) -> Tuple[torch.Tensor]:
         cur_len = input_ids.shape[-1]
         batch_size = len(self._beam_hyps)
-        assert batch_size == (input_ids.shape[0] // self.group_size)
+        if not (batch_size == (input_ids.shape[0] // self.group_size)):
+            if self.num_beam_groups > 1:
+                raise ValueError(
+                    f"A group beam size of {input_ids.shape[0]} is used as the input, but a group beam "
+                    f"size of {self.group_size} is expected by the beam scorer."
+                )
+            else:
+                raise ValueError(
+                    f"A beam size of {input_ids.shape[0]} is used as the input, but a beam size of "
+                    f"{self.group_size} is expected by the beam scorer."
+                )
 
         device = input_ids.device
         next_beam_scores = torch.zeros((batch_size, self.group_size), dtype=next_scores.dtype, device=device)
@@ -223,12 +233,10 @@ class BeamSearchScorer(BeamScorer):
 
         for batch_idx, beam_hyp in enumerate(self._beam_hyps):
             if self._done[batch_idx]:
-                assert (
-                    len(beam_hyp) >= self.num_beams
-                ), f"Batch can only be done if at least {self.num_beams} beams have been generated"
-                assert (
-                    eos_token_id is not None and pad_token_id is not None
-                ), "generated beams >= num_beams -> eos_token_id and pad_token have to be defined"
+                if self.num_beams < len(beam_hyp):
+                    raise ValueError(f"Batch can only be done if at least {self.num_beams} beams have been generated")
+                if eos_token_id is None or pad_token_id is None:
+                    raise ValueError("Generated beams >= num_beams -> eos_token_id and pad_token have to be defined")
                 # pad the batch
                 next_beam_scores[batch_idx, :] = 0
                 next_beam_tokens[batch_idx, :] = pad_token_id