🐛 fix(llava_onevision): use isinstance for type checking of visual input

EvolvingLMMs-Lab · Dec 22, 2024 · 67b582e · 67b582e
1 parent 7f9c482
commit 67b582e
Showing 1 changed file with 2 additions and 1 deletion.
diff --git a/lmms_eval/models/llava_onevision.py b/lmms_eval/models/llava_onevision.py
@@ -279,7 +279,8 @@ def loglikelihood(self, requests: List[Instance]) -> List[Tuple[float, bool]]:
 
                     task_type = "video"
 
-                elif type(visual[0]) == PIL.Image.Image:
+                # elif type(visual[0]) == PIL.Image.Image:
+                elif isinstance(visual[0], PIL.Image.Image):
                     image_tensor = process_images(visual, self._image_processor, self._config)
                     if type(image_tensor) is list:
                         image_tensor = [_image.to(dtype=torch.float16, device=self.device) for _image in image_tensor]