Update app.py

app.py

@@ -41,10 +41,10 @@ def process_video_audio(video_path, audio_path):
                 (x, y, w, h) = faces[0]
                 face = frame[y:y+h, x:x+w]
                 resized_face = cv2.resize(face, (120, 120))
-                train_visual[0, :, :, :, frame_idx] = resized_face
+                train_visual[0, :, :, :, frame_idx] = pt.tensor(resized_face)
             else:
                 resized_frame = cv2.resize(frame, (120, 120))
-                train_visual[0, :, :, :, frame_idx] = resized_frame
+                train_visual[0, :, :, :, frame_idx] = pt.tensor(resized_frame)
             last_frame = frame
             frame_idx += 1
     cap.release()