Update app.py

app.py

@@ -30,6 +30,13 @@ def audio_to_text(audio, num_speakers):
   if duration > 4 * 60 * 60:
     return "Audio duration too long"
 
+
+  model = whisper.load_model("large-v2")
+  embedding_model = PretrainedSpeakerEmbedding( 
+    "speechbrain/spkrec-ecapa-voxceleb",
+    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+    )
+
   result = model.transcribe(path)
   segments = result["segments"]