VidLA: Video-Language Alignment at Scale | Read Paper on Bytez