Towards More Unified In-context Visual Understanding | Read Paper on Bytez