FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification

Devs

FasterDiT: Towards Faster Diffusion Transformers Training without Architecture Modification | Read Paper on Bytez