Multimodal model

TinyMU

TinyMU is a compact 229M audio-language model for music understanding and reasoning. It is designed to answer music-related questions efficiently, using a small architecture that targets strong performance under tight compute budgets instead of scaling to large audio-language model size.