Update app.py

app.py

@@ -142,7 +142,7 @@ def model_inference(input_dict, history):
         images=images if images else None,
         return_tensors="pt",
         padding=True,
-    ).to("cuda")
+    ).to("cpu")
     streamer = TextIteratorStreamer(processor, skip_prompt=True, skip_special_tokens=True)
     generation_kwargs = dict(inputs, streamer=streamer, max_new_tokens=1024)
     thread = Thread(target=model.generate, kwargs=generation_kwargs)