Reactions
3 postsSH 
Sumanth Hegde@sumanthrhExcited 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…
- likes: 100
- replies: 3
- bookmarks: 71
- reposts: 20
VL vLLM@vllm_projectThis weight transfer engine is native to vLLM. Any Ray-based trainer can adopt it with a single WeightSource iterator. Great work here! 🔗
- likes: 143
- replies: 1
- bookmarks: 67
- reposts: 19
VL vLLM@vllm_projectThis weight transfer engine is native to vLLM. Any Ray-based trainer can adopt it with a single WeightSource iterator. Great work here! 🔗
- likes: 143
- replies: 1
- bookmarks: 67
- reposts: 19
