IsoExec unifies numerical execution across SkyRL's vLLM and Megatron runtimes, reducing the average rollout-versus-training logprob difference below 1e-6 on Qwen3.5-35B-A3B with 25% overhead.
IsoExec: Unified Execution to Eliminate Trainer-Inference Mismatch in SkyRL
Source:vLLM Blog ↗