Split-KV: Efficient Context Parallelism for Training Diffusion Drafters

arXiv:2611.05926 · cs.AI · Submitted 2026-11-04 · Read on arXiv

cs.AI

Submitted: 2026-11-04

Updated: 2026-11-04

Code: https://github.com/zxgx/verl

Terminology

Sources

Related papers