Update app.py

app.py

@@ -246,8 +246,8 @@ def generate(
         return
 
     # New branch for video processing with Qwen2VL.
-    if lower_text.startswith("@qwen2vl-video"):
-        prompt = text[len("@qwen2vl-video"):].strip()
+    if lower_text.startswith("@video-infer"):
+        prompt = text[len("@video-infer"):].strip()
         if files:
             # Assume the first file is a video.
             video_path = files[0]
@@ -376,9 +376,9 @@ demo = gr.ChatInterface(
         gr.Slider(label="Repetition penalty", minimum=1.0, maximum=2.0, step=0.05, value=1.2),
     ],
     examples=[
-        [{"text": "@qwen2vl-video Describe the Ad", "files": ["examples/coca.mp4"]}],
-        [{"text": "@qwen2vl-video Summarize the event in video", "files": ["examples/sky.mp4"]}],
-        [{"text": "@qwen2vl-video Describe the video", "files": ["examples/Missing.mp4"]}],
+        [{"text": "@video-infer Describe the Ad", "files": ["examples/coca.mp4"]}],
+        [{"text": "@video-infer Summarize the event in video", "files": ["examples/sky.mp4"]}],
+        [{"text": "@video-infer Describe the video", "files": ["examples/Missing.mp4"]}],
         ["@image Chocolate dripping from a donut"],
         ["Python Program for Array Rotation"],
         ["@tts1 Who is Nikola Tesla, and why did he die?"],
@@ -388,9 +388,9 @@ demo = gr.ChatInterface(
     ],
     cache_examples=False,
     type="messages",
-    description="# **QwQ Edge `@qwen2vl-video 'prompt..', @image, @tts1`**",
+    description="# **QwQ Edge `@video-infer 'prompt..', @image, @tts1`**",
     fill_height=True,
-    textbox=gr.MultimodalTextbox(label="Query Input", file_types=["image", "video"], file_count="multiple", placeholder="‎ @tts1, @tts2-voices, @image for image gen, @qwen2vl-video for video, default [text, vision]"),
+    textbox=gr.MultimodalTextbox(label="Query Input", file_types=["image", "video"], file_count="multiple", placeholder="‎ @tts1, @tts2-voices, @image for image gen, @video-infer for video, default [text, vision]"),
     stop_btn="Stop Generation",
     multimodal=True,
 )