Abstract: This paper investigates the zero-shot capabilities of pre-trained large language models (LLMs) for music genre classification. The proposed approach splits audio signals into 20 ms chunks ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results