Model techniques map
Techniquespost-trainingrollout & RL infrastructure

general family · filed under post-training

Asynchronous reinforcement learning framework

A framework for asynchronous RL in which rollout phases may be interrupted to switch to updated policy checkpoints.

Also called asynchronous RL framework.

source
1
model
1
lab adopt it
1
strongest
core

How sources treat it

One count per evidence span, weakest treatment to strongest.

core 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

In our asynchronous RL framework, the rollout phase is periodically interrupted to switch to updated policy checkpoints.

corepost trainingin DeepSeek-V4.1-FlashDeepSeek

Filed alongside

Other methods under post-training :: rollout & RL infrastructure.