specific method · filed under post-training
Asynchronous Group Relative Policy Optimization
Group Relative Policy Optimization run fully asynchronously, including on very large batches.
- sources
- 2
- models
- 2
- lab adopt it
- 1
- strongest
- core
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1core 1
Documented in
Evidence
2 spans quoted from the sources, strongest treatment first.
Fully asynchronous Group Relative Policy Optimization (GRPO) on very large batches
coreoptimizationin MiMo-V2.6-Flash-RLXiaomi
Fully asynchronous Group Relative Policy Optimization (GRPO) on very large batches
usedoptimizationin MiMo-V2.6-Pro-RLXiaomi
Filed alongside
Other methods under post-training :: reinforcement learning algorithm.
Group Relative Policy OptimizationReinforcement LearningGroupwise Advantage RedistributionFreezing the MoE Router During Reinforcement LearningIcePopOff-Policy Sequence MaskingReinforcement Learning from Verifiable RewardsRollout Routing ReplayUnbiased KL EstimateChain-of-Thought Reinforcement LearningDirect Double-Sided Importance SamplingKeep Sampling MaskMixed Reinforcement LearningPivot Reinforcement LearningReinforcement Learning Post-TrainingAbstention TrainingAdvantage ShapingAgentic RL Task MixCISPO with Length-Weighted Leave-One-Out Group-Relative AdvantagesConcatenated Routing ReplayDerived-Latency PenaltyDomain-Specialized RL ExpertsDomain-Specific GRPO TrainingDropping All-Zero-Advantage Groups