Sumanth Hegde(@sumanthrh):Excited to share our work on large-scale sharded weight transfer! We’ve implemented a native sharded weight transfer engine in vLLM and SkyRL using Ray Direct Transfer (RDT) + NIXL. We’re able to achieve weight transfer for Kimi K2 (1T params) in BF16 in 7.53 seconds on 48 8xH100 nodes. Blog:

Sumanth Hegde

@sumanthrh

Post-training @anyscalecompute. Prev - @UCSanDiego, @C3_AI, @iitmadras. Machine Learning and Systems. Intensity is all you need.

加入 February 2016

26 正在关注 1.2K 粉丝

Sumanth Hegde@sumanthrh

2026.08.25 17:06

Excited to share our work on large-scale sharded weight transfer! We’ve implemented a native sharded weight transfer engine in vLLM and SkyRL using Ray Direct Transfer (RDT) + NIXL. We’re able to achieve weight transfer for Kimi K2 (1T params) in BF16 in 7.53 seconds on 48 8xH100 nodes. Blog:

显示更多