Update app.py

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 import numpy as np
 import torch
-from transformers import AutoProcessor, pipeline, BarkModel, GenerationConfig
+from transformers import AutoProcessor, pipeline, BarkModel
 
 ASR_MODEL_NAME = "bofenghuang/whisper-large-v2-cv11-german"
 TTS_MODEL_NAME = "suno/bark-small"
@@ -16,16 +16,14 @@ device = "cuda:0" if torch.cuda.is_available() else "cpu"
 # load speech translation checkpoint
 asr_pipe = pipeline("automatic-speech-recognition", model=ASR_MODEL_NAME, chunk_length_s=10,device=device)
 
-# update the generation config
-MULTILINGUAL = True  # set True for multilingual models, False for English-only
-generation_config = GenerationConfig.from_pretrained("openai/whisper-large-v2")
-
-
 # load text-to-speech checkpoint
 processor = AutoProcessor.from_pretrained("suno/bark-small")
 model = BarkModel.from_pretrained("suno/bark-small").to(device)
 sampling_rate = model.generation_config.sample_rate
 
+# set the forced ids
+model.config.forced_decoder_ids = processor.get_decoder_prompt_ids(task="translate")
+
 def translate(audio):
     outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "translate"})
     return outputs["text"]