diff --git a/modelscope/outputs.py b/modelscope/outputs.py
index 30860f29..0937e441 100644
--- a/modelscope/outputs.py
+++ b/modelscope/outputs.py
@@ -159,10 +159,17 @@ TASK_OUTPUTS = {
     # }
     Tasks.action_recognition: [OutputKeys.LABELS],
 
+    # live category recognition result for single video
+    # {
+    #       "scores": [0.885272, 0.014790631, 0.014558001],
+    #       'labels': ['修身型棉衣', '高腰牛仔裤', '休闲连体裤']
+    # }
+    Tasks.live_category: [OutputKeys.SCORES, OutputKeys.LABELS],
+
     # video category recognition result for single video
     # {
-    #       "scores": [0.7716429233551025]
-    #       "labels": ['生活>>好物推荐'],
+    #       "scores": [0.7716429233551025],
+    #       "labels": ['生活>>好物推荐']
     # }
     Tasks.video_category: [OutputKeys.SCORES, OutputKeys.LABELS],
 
diff --git a/modelscope/pipelines/cv/live_category_pipeline.py b/modelscope/pipelines/cv/live_category_pipeline.py
index b7f3202e..c16ba6ba 100644
--- a/modelscope/pipelines/cv/live_category_pipeline.py
+++ b/modelscope/pipelines/cv/live_category_pipeline.py
@@ -85,7 +85,7 @@ class LiveCategoryPipeline(Pipeline):
             for label_key in label_keys:
                 if label_info[label_key] not in label_str:
                     label_str.append(label_info[label_key])
-            labels.append('>>'.join(label_str))
+            labels.append(label_str[-1])
         return {OutputKeys.SCORES: list(scores), OutputKeys.LABELS: labels}
 
     def postprocess(self, inputs: Dict[str, Any]) -> Dict[str, Any]:
diff --git a/modelscope/utils/constant.py b/modelscope/utils/constant.py
index ef3c4f4f..4d4ad9e9 100644
--- a/modelscope/utils/constant.py
+++ b/modelscope/utils/constant.py
@@ -61,7 +61,6 @@ class CVTasks(object):
     video_category = 'video-category'
 
     video_embedding = 'video-embedding'
-
     virtual_try_on = 'virtual-try-on'