Upload infer_qwen2_vl.py with huggingface_hub

infer_qwen2_vl.py

@@ -78,12 +78,13 @@ for batch_idx in tqdm(range(begin, end, batch_size)):
             messages, tokenize=False, add_generation_prompt=True
         )
         image_inputs, video_inputs = process_vision_info(messages)
+        print(video_inputs)
         image_inputs_batch.append(image_inputs)
         video_inputs_batch.append(video_inputs)
     inputs = processor(
         text=[text],
         images=image_inputs_batch,
-        videos=video_inputs_batch,
+        videos=None,
         padding=True,
         return_tensors="pt",
     )