diff --git a/src/llamafactory/data/mm_plugin.py b/src/llamafactory/data/mm_plugin.py index f8598212..40378f45 100644 --- a/src/llamafactory/data/mm_plugin.py +++ b/src/llamafactory/data/mm_plugin.py @@ -525,14 +525,14 @@ class InternVLPlugin(BasePlugin): mm_inputs = {} image_video_patches = [] - if len(images) != 0 and isinstance(images[0], str): + if len(images) != 0: images = self._regularize_images( images, image_max_pixels=getattr(processor, "image_max_pixels", 1024 * 1024), image_min_pixels=getattr(processor, "image_min_pixels", 32 * 32), )["images"] - if len(videos) != 0 and isinstance(videos[0], str): + if len(videos) != 0: videos = self._regularize_videos( videos, image_max_pixels=getattr(processor, "video_max_pixels", 256 * 256),