[Bugfix] Fix BAGEL online serving for text and image understanding (#31546)

Signed-off-by: Dylan1229 <yvanphys@gmail.com>
Signed-off-by: UED <zxr3611244710@gmail.com>
Signed-off-by: mr-ye-cao <yecaoyc2019@gmail.com>
Co-authored-by: UED <zxr3611244710@gmail.com>
Co-authored-by: mr-ye-cao <yecaoyc2019@gmail.com>
Co-authored-by: Mr-Ye-Cao <60802056+Mr-Ye-Cao@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

vllm/model_executor/models/bagel.py

@@ -346,6 +346,13 @@ class BagelForConditionalGeneration(
         }
     )
 
+    @classmethod
+    def get_placeholder_str(cls, modality: str, i: int) -> str | None:
+        if modality.startswith("image"):
+            return "<|image_pad|>"
+
+        raise ValueError("Only image modality is supported")
+
     def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         super().__init__()