Update app.py

app.py

@@ -28,7 +28,8 @@ def transcribe(audio):
     sr, y = audio
     y = y.astype(np.float32)
     y /= np.max(np.abs(y))
-    return pipeline({"sampling_rate": sr, "raw": y.astype(np.half)}, generate_kwargs={"forced_decoder_ids": forced_decoder_ids}, max_new_tokens=255)["text"]
+    with torch.autocast("cuda"):
+        return pipeline({"sampling_rate": sr, "raw": y}, generate_kwargs={"forced_decoder_ids": forced_decoder_ids}, max_new_tokens=255)["text"]
 
 examples = [
     "./examples/audio_1.mp3",