solve problem with input

app.py

@@ -14,10 +14,10 @@ transform = midas_transforms.small_transform
 def predict_depth(image):
     img = np.array(image)
     img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
-    input_tensor = transform(img_rgb).unsqueeze(0)
+    input_tensor = transform(img_rgb)
 
     with torch.no_grad():
-        prediction = midas(input_tensor)
+        prediction = midas(input_tensor.unsqueeze(0))
         prediction = torch.nn.functional.interpolate(
             prediction.unsqueeze(1),
             size=img_rgb.shape[:2],