tried to fix error

app.py

@@ -4,44 +4,32 @@ import spaces
 import torch
 from transformers import BlipProcessor, BlipForConditionalGeneration
 
-
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 processor = BlipProcessor.from_pretrained("noamrot/FuseCap_Image_Captioning")
 model = BlipForConditionalGeneration.from_pretrained("noamrot/FuseCap_Image_Captioning").to(device)
 
 @spaces.GPU(duration=15)
 def inference(raw_image):
-    global model
-    cur_device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    model = model.to(device)
     text = "a picture of "
-    inputs = processor(raw_image, text, return_tensors="pt").to(cur_device)
+    inputs = processor(raw_image, text, return_tensors="pt").to(device)
     out = model.generate(**inputs)
     caption = processor.decode(out[0], skip_special_tokens=True)
     return caption
 
-
-inputs = [gr.Image(type='pil', interactive=False),]
-# outputs = gr.outputs.Textbox(label="Caption")
+inputs = gr.Image(type="pil", interactive=False)
 outputs = gr.Textbox(label="Caption")
 
 description = "Gradio demo for FuseCap: Leveraging Large Language Models for Enriched Fused Image Captions. This demo features a BLIP-based model, trained using FuseCap."
 examples = [["surfer.jpg"], ["bike.jpg"]]
 article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2305.17718' target='_blank'>FuseCap: Leveraging Large Language Models for Enriched Fused Image Captions</a>"
 
-
 iface = gr.Interface(fn=inference, 
-                    inputs="image",
-                    outputs="text", 
-                    title="FuseCap",
-                    description=description,
-                    article=article,
-                    examples=examples,
-                    # enable_queue=True
-                    )
+                     inputs=inputs,
+                     outputs=outputs, 
+                     title="FuseCap",
+                     description=description,
+                     article=article,
+                     examples=examples)
 
 iface.queue()
-
-
 iface.launch()
-