requirements

app.py

@@ -617,7 +617,7 @@ def sample(
     return complete_video
 
 
-@spaces.GPU(duration=200)
+@spaces.GPU(duration=150)
 @torch.no_grad()
 def process_video(video_input, audio_input, max_num_seconds):
     """Main processing function to generate synchronized video"""
@@ -945,29 +945,29 @@ def process_video(video_input, audio_input, max_num_seconds):
     #     return None
 
 
-def get_max_duration(video_input, audio_input):
-    """Get the maximum duration in seconds for the slider"""
-    try:
-        # Default to 60 seconds if files don't exist
-        if video_input is None or not os.path.exists(video_input):
-            video_input = DEFAULT_VIDEO_PATH
+# def get_max_duration(video_input, audio_input):
+#     """Get the maximum duration in seconds for the slider"""
+#     try:
+#         # Default to 60 seconds if files don't exist
+#         if video_input is None or not os.path.exists(video_input):
+#             video_input = DEFAULT_VIDEO_PATH
 
-        if audio_input is None or not os.path.exists(audio_input):
-            audio_input = DEFAULT_AUDIO_PATH
+#         if audio_input is None or not os.path.exists(audio_input):
+#             audio_input = DEFAULT_AUDIO_PATH
 
-        # Get video duration
-        video_reader = decord.VideoReader(video_input)
-        video_duration = len(video_reader) / video_reader.get_avg_fps()
+#         # Get video duration
+#         video_reader = decord.VideoReader(video_input)
+#         video_duration = len(video_reader) / video_reader.get_avg_fps()
 
-        # Get audio duration
-        raw_audio = get_raw_audio(audio_input, 16000)
-        audio_duration = len(raw_audio) / 25  # Assuming 25 fps
+#         # Get audio duration
+#         raw_audio = get_raw_audio(audio_input, 16000)
+#         audio_duration = len(raw_audio) / 25  # Assuming 25 fps
 
-        # Return the minimum of the two durations
-        return min(video_duration, audio_duration)
-    except Exception as e:
-        print(f"Error getting max duration: {str(e)}")
-        return 60  # Default to 60 seconds
+#         # Return the minimum of the two durations
+#         return min(video_duration, audio_duration)
+#     except Exception as e:
+#         print(f"Error getting max duration: {str(e)}")
+#         return 60  # Default to 60 seconds
 
 
 def new_game_click(state):
@@ -1015,38 +1015,38 @@ with gr.Blocks(
                         else None,
                     )
 
-                    max_duration = gr.State(value=60)  # Default max duration
+                    # max_duration = gr.State(value=60)  # Default max duration
 
-                    max_seconds_slider = gr.Slider(
-                        minimum=0,
-                        maximum=60,  # Will be updated dynamically
-                        value=0,
-                        step=1,
-                        label="Max Duration (seconds, 0 = full length)",
-                        info="Limit the processing duration (0 means use full length)",
-                    )
+                    # max_seconds_slider = gr.Slider(
+                    #     minimum=0,
+                    #     maximum=60,  # Will be updated dynamically
+                    #     value=0,
+                    #     step=1,
+                    #     label="Max Duration (seconds, 0 = full length)",
+                    #     info="Limit the processing duration (0 means use full length)",
+                    # )
 
                     process_button = gr.Button("Generate Synchronized Video")
 
                 with gr.Column("Output Video"):
                     video_output = gr.Video(label="Output Video", width=512, height=512)
 
-            # Update slider max value when inputs change
-            def update_slider_max(video, audio):
-                max_dur = get_max_duration(video, audio)
-                return {"maximum": max_dur, "__type__": "update"}
+            # # Update slider max value when inputs change
+            # def update_slider_max(video, audio):
+            #     max_dur = get_max_duration(video, audio)
+            #     return {"maximum": max_dur, "__type__": "update"}
 
-            video_input.change(
-                update_slider_max, [video_input, audio_input], [max_seconds_slider]
-            )
-            audio_input.change(
-                update_slider_max, [video_input, audio_input], [max_seconds_slider]
-            )
+            # video_input.change(
+            #     update_slider_max, [video_input, audio_input], [max_seconds_slider]
+            # )
+            # audio_input.change(
+            #     update_slider_max, [video_input, audio_input], [max_seconds_slider]
+            # )
 
             # Show Wordle message when processing starts and hide when complete
             process_button.click(
                 fn=process_video,
-                inputs=[video_input, audio_input, max_seconds_slider],
+                inputs=[video_input, audio_input, 10],
                 outputs=video_output,
             )