change to cpu

app.py

@@ -10,7 +10,7 @@ torch.set_float32_matmul_precision(["high", "highest"][0])
 birefnet = AutoModelForImageSegmentation.from_pretrained(
     "ZhengPeng7/BiRefNet", trust_remote_code=True
 )
-birefnet.to("cuda")
+birefnet.to("cpu")
 
 transform_image = transforms.Compose(
     [
@@ -27,10 +27,10 @@ def fn(image):
     processed_image = process(im)
     return (processed_image, origin)
 
-@spaces.GPU
+#@spaces.GPU
 def process(image):
     image_size = image.size
-    input_images = transform_image(image).unsqueeze(0).to("cuda")
+    input_images = transform_image(image).unsqueeze(0).to("cpu")
     # Prediction
     with torch.no_grad():
         preds = birefnet(input_images)[-1].sigmoid().cpu()