specific method · filed under post-training
Co-located RL training
Places rollout and training on the same physical devices, time-sharing their execution to manage compute resources.
- sources
- 2
- models
- 2
- labs adopt it
- 2
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 2
Documented in
Evidence
2 spans quoted from the sources, strongest treatment first.
We adopt co-located RL training [58] to keep each 1M-context Kimi K3 RL experiment within a few hundred GPUs
usedpost trainingin Kimi K3Moonshot AI
We colocate rollout and training on the same physical devices and time-share their execution, eliminating the need to manually tune resource allocation between the two phases.
usedpost trainingin DeepSeek-V4.1DeepSeek
Filed alongside
Other methods under post-training :: rollout & RL infrastructure.
Asynchronous reinforcement learningPartial rolloutAsynchronous RL frameworks for large-scale agent scaffolds and environment orchestrationToken-in-token-out (TITO)Asynchronous reinforcement learning infrastructureData SchedulerDecoupled control plane and data planeDecoupling agent rollout into sandbox and worker containerDeficit-corrected schedulingLarge-scale asynchronous RL in synthesized tasksOne-step off-policy asynchronous reinforcement learningPredictive Rollout DispatchSample-grained garbage collectionSeamless Rollout EngineSLIMEToken-granularity persistence of rollout statesToken-level interruptionTool ManagerToolboxAdaptive Rollout ConcurrencyAdaptive Rollout SchedulingAgent LoopAgent-centric rollout executionAsynchronous Agent RL algorithms