Update app.py

app.py

@@ -80,7 +80,8 @@ def inference_point(input_img, evt: gr.SelectData,):
     
     # Assume 'points' is a list of (x, y) coordinates to specify where the user clicks
     # Process the image and points to create a segmentation map accordingly
-    _, visualized_output = demo.run_on_image_with_points(img, points)
+    text_features = torch.from_numpy(np.load("./text_embedding/lvis_coco_text_embedding.npy"))
+    _, visualized_output = demo.run_on_image_with_points(img, points,text_features)
     end_time = time.time()
     print("inf time",end_time - start_time)
     return visualized_output