Ground-V: Teaching VLMs to Ground Complex Instructions in Pixels | Read Paper on Bytez