Hacker News new | ask | show | jobs
by GaggiX 871 days ago
Well for LLaVa-1.6 we know it was trained only on English multi-modal data.