Long Video Diffusion Generation with Segmented Cross-Attention and Content-Rich Video Data Curation | Read Paper on Bytez