1 paper · 1 filter
Yucheng Guo, Yongjian Guo, Zhong Guan +6
In video generation models, particularly world models, training large-scale video diffusion Transformers (such as DiT and MMDiT) poses significant computational challenges due to t…