Update app.py

app.py

@@ -46,17 +46,21 @@ def inference(
     else:
         voice_samples, conditioning_latents = load_voices(voices)
 
-    start_time = time.time()
+    audio_frames = []
 
     for j, text in enumerate(texts):
         for audio_frame in tts.tts_with_preset(
             text,
             voice_samples=voice_samples,
             conditioning_latents=conditioning_latents,
-            preset= Preset,
+            preset=Preset,
             k=1
         ):
-            yield (24000, audio_frame.cpu().detach().numpy())
+            audio_frames.append(torch.from_numpy(audio_frame.cpu().detach().numpy())) 
+
+    complete_audio = torch.cat(audio_frames, dim=0)
+
+    yield (24000, complete_audio.numpy()) 
 
 def main():
     title = "Tortoise TTS "