Model techniques map
Techniquespost-trainingrollout & RL infrastructure

implementation detail · filed under post-training

Continuous rollout with asynchronous reward computation and early termination

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

source
1
models
2
lab adopt it
1
strongest
used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

Our design integrates continuous rollout, asynchronous reward computation, and early termination to minimize GPU idle time

usedsoftware implementationin MiMo-7B series RL training infrastructureXiaomi

Filed alongside

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