Update app.py

app.py

@@ -5,7 +5,7 @@ from lavis.models import load_model_and_preprocess
 
 
 # Load the Blip-Caption model
-device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
 model, vis_processors, _ = load_model_and_preprocess(name="blip_caption", model_type="base_coco", is_eval=True, device=device)
 
 # Define the input and output functions for Gradio
@@ -14,7 +14,7 @@ def generate_caption(image_file):
         raw_image = Image.open(image_file).convert("RGB")
 
         # Preprocess the image using the Blip-Caption model's visual processors
-        image = vis_processors["eval"](raw_image).unsqueeze(0).to(device)
+        image = vis_processors["eval"](raw_image).unsqueeze(0)
 
         # Generate captions using the Blip-Caption model
         captions = model.generate({"image": image}, use_nucleus_sampling=True, num_captions=5)