The ANWSER is just REPEAT the question in audio.

[sfofgalaxy]

Why do the model only repeat the question in my audio and do not answer the question in my audio? I put prompt here:

prompt_template.format(" " + (prompt.strip() if prompt!=None and prompt!="" else "Please answer the speaker's question."))