Increase Flexibility for OOV Multimodal Token Handling (#34858)

Signed-off-by: Alex Brooks <albrooks@redhat.com>

vllm/model_executor/models/molmo2.py

@@ -2711,13 +2711,11 @@ class Molmo2ForConditionalGeneration(
         multimodal_embeddings: MultiModalEmbeddings | None = None,
         *,
         is_multimodal: torch.Tensor | None = None,
-        handle_oov_mm_token: bool = False,
     ) -> torch.Tensor:
         inputs_embeds = self._embed_text_input_ids(
             input_ids,
             self.get_language_model().embed_input_ids,
             is_multimodal=is_multimodal,
-            handle_oov_mm_token=handle_oov_mm_token,
         )
 
         if multimodal_embeddings is None or len(multimodal_embeddings) == 0: