Only giving Hindi Transcription, even after specifying language

[nikhil131102]

transcript = pipeline('/content/kannada.mp3',task='transcribe')
print(transcript)

Pseudocode - adapt based on library support

forced_decoder_ids = pipeline.tokenizer.get_decoder_prompt_ids(language="kn", task="transcribe")
pipeline.model.config.forced_decoder_ids = forced_decoder_ids

transcript = pipeline('/content/kannada.mp3', task='transcribe')
print(transcript)