Update app.py

app.py

@@ -144,6 +144,8 @@ def get_caption(img,model_name):
     model = model.eval() 
     model = model.to(device)
 
+    clip_model = clip_model.to(device)
+
     input = clip_processor(images=img, return_tensors="pt").to(device)
     with torch.no_grad():
         prefix = clip_model.get_image_features(**input)