Open Models Almanac
Methodspost-trainingrollout & RL infrastructure

single mention · post-training

Continuous rollout with asynchronous rewards and early termination

Combines continuous rollout, asynchronous reward computation, and early termination to reduce GPU idle time.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See rollout & RL infrastructure for the methods filed beside it.

Filed under post-training › rollout & RL infrastructure