[Bugfix][v1] fixed llava-hf/llava-1.5-7b-hf is broken on V1 (#14554)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/model_executor/models/llava.py

@@ -783,9 +783,13 @@ class LlavaForConditionalGeneration(nn.Module, SupportsMultiModal, SupportsPP):
         if image_input is None:
             return None
         vision_embeddings = self._process_image_input(image_input)
-        if kwargs.get("v0_path", False):
+
+        if kwargs.get("v0_path", False) or \
+            image_input.get("feat_is_patch") is None or \
+                image_input.get("embed_is_patch") is None:
+            # The path is used for pixtral (V0 only) and llava (V0/V1)
             return vision_embeddings
-        else:
+
         nested_emb = [
             self._get_mm_embeds(*args) for args in zip(
                 vision_embeddings, image_input["feat_is_patch"],