How Far Can Chord-Symbol Time-Series Adaptation Carry Genre Identity? Capabilities and Boundaries in Multi-Genre Chord-Symbol Modeling

A frozen pop-jazz Music Transformer adapted via five PEFT methods (LoRA, IA3, BitFit, prefix tuning, full fine-tuning) across 11 genres shows macro chord-prediction accuracy gains of +2.89 to +3.6 percentage points over the frozen baseline across a 165-cell evaluation grid. All adaptation methods improve genre-local harmonic modeling, with chord-symbol sequences treated as interpretable time series rather than complete musical representations.