fix

app.py

@@ -20,7 +20,7 @@ def snap(image, video):
 
 demo = gr.Interface(
     snap,
-    inputs=[gr.Image(source="webcam", tool=None),
+    inputs=[gr.Image(source="webcam", tool=None, label="Input Image", type="pil"),
             gr.Video(source="webcam")],
     outputs=[gr.Image(label="RGB"),
              gr.Image(label="predicted depth"),

inference.py

@@ -47,6 +47,7 @@ class SegmentPredictor:
         self.mask_generator = SamAutomaticMaskGenerator(sam)
     
     def predict(self, image):
+        image = np.array(image)
         sam_result = self.mask_generator.generate(image)
         mask_annotator = sv.MaskAnnotator()
         detections = sv.Detections.from_sam(sam_result=sam_result)