vLLM and SkyRL add day-0 RL support for Beam

AISummary

vLLM says SkyRL now supports RL training for Beam with LoRA, BF16 full fine-tuning, and MXFP8 RL, using vLLM rollouts and Reflection's Megatron-Bridge implementation.

SkyRL reports a 62% LoRA decode throughput gain from right-sizing expert adapters in vLLM, rising from 276 to 447 tok/s/GPU at 32 concurrent requests on PP2×TP8, though this is a decode result rather than an end-to-end training speedup.

Post on XView on X

vLLM

@vllm_project

Part of a thread · earlier post

🎉 Day-0 RL support for Beam in SkyRL.

LoRA, BF16 full fine-tuning and MXFP8 RL, with vLLM rollouts and Reflection’s Megatron-Bridge implementation.

SkyRL reports +62% LoRA decode throughput from right-sizing expert adapters in vLLM: 276 → 447 tok/s/GPU at 32 concurrent requests, PP2×TP8. This is a decode result, not an end-to-end training speedup.

R3 works with pipeline parallelism, and FP32 router weights are preserved during weight sync. Thanks to Sumanth Hegde and the SkyRL team!

https://docs.skyrl.ai/

5/6

Source: vLLM · x.comPublished Oct 16, 9:23 AM ET · added here 9:34 AM ET

vLLM and SkyRL add day-0 RL support for Beam

Topics