Update app.py

app.py

@@ -4,6 +4,7 @@ import torch
 from datasets import load_dataset
 
 from transformers import SpeechT5ForTextToSpeech, SpeechT5HifiGan, SpeechT5Processor, pipeline
+from transformers import VitsModel, VitsTokenizer
 
 
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
@@ -21,8 +22,11 @@ embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validat
 speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
 
 
+model = VitsModel.from_pretrained("Matthijs/mms-tts-deu").to(device)
+tokenizer = VitsTokenizer.from_pretrained("Matthijs/mms-tts-deu")
+
 def translate(audio):
-    outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "translate"})
+    outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "transcribe", "language": "de"})
     return outputs["text"]