Upload 2 files

app.py

@@ -112,7 +112,8 @@ def load_models():
 
 def preprocess_audio(audio_path):
     """Convert audio to 16kHz mono"""
-    waveform, sr = torchaudio.load(audio_path)
+    # Use soundfile backend to avoid torchcodec dependency
+    waveform, sr = torchaudio.load(audio_path, backend="soundfile")
     
     # Convert to mono
     if waveform.shape[0] > 1:

requirements.txt

@@ -7,3 +7,4 @@ transformers>=4.30.0
 gradio>=4.0.0
 openai-whisper>=20231117
 numpy>=1.24.0
+soundfile>=0.12.0