I’m working on something similar - ASR for dysarthric speech - and I’ll happily follow your project so maybe we can both benefit. Do you have some X/Mastodon account where you share the progress?