3 ms·Well for LLaVa-1.6 we know it was trained only on English multi-modal data.by GaggiX 3y agoWell for LLaVa-1.6 we know it was trained only on English multi-modal data.