Last released Oct 12, 2026
An easy-to-use, fast toolkit to scale up RL post-training on a single node.