init

.gitignore

@@ -180,3 +180,4 @@ temp/
 src/moviedubber/infer/basic_test.toml
 
 upload.py
+output

app.py

@@ -73,6 +73,10 @@ asr_pipe = load_asr_model()
 
 videofeature_extractor = VideoFeatureExtractor(device=device)
 
+out_dir = "./output"
+if not os.path.exists(out_dir):
+    os.makedirs(out_dir)
+
 
 def deepdubber(video_path: str, subtitle_text: str, audio_path: str = None) -> str:
     pixel_values, num_patches_list = load_video(video_path, num_segments=8, max_num=1)
@@ -177,7 +181,7 @@ def deepdubber(video_path: str, subtitle_text: str, audio_path: str = None) -> s
         generated_wave = generated_wave.squeeze().cpu().numpy()
 
     # using a temporary wav file to save the generated audio
-    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_wav_file:
+    with tempfile.NamedTemporaryFile(delete=False, suffix=".wav", dir="./output") as temp_wav_file:
         temp_wav_path = temp_wav_file.name
         soundfile.write(temp_wav_path, generated_wave, samplerate=24000)
 
@@ -262,4 +266,4 @@ def create_ui():
 
 if __name__ == "__main__":
     app = create_ui()
-    app.launch()
+    app.launch(allowed_paths=["./output"])