AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Training
read the original abstract
Reinforcement learning (RL) has become a pivotal technology in the post-training phase of large language models (LLMs). Traditional task-colocated RL frameworks suffer from significant scalability bottlenecks, while task-separated RL frameworks face challenges in complex dataflows and the corresponding resource idling and workload imbalance. Moreover, most existing frameworks are tightly coupled with LLM training or inference engines, making it difficult to support custom-designed engines. To address these challenges, we propose AsyncFlow, an asynchronous streaming RL framework for efficient post-training. Specifically, we introduce a distributed data storage and transfer module that provides a unified data management and fine-grained scheduling capability in a fully streamed manner. This architecture inherently facilitates automated pipeline overlapping among RL tasks and dynamic load balancing. Moreover, we propose a producer-consumer-based asynchronous workflow engineered to minimize computational idleness by strategically deferring parameter update process within staleness thresholds. Finally, the core capability of AsynFlow is architecturally decoupled from underlying training and inference engines and encapsulated by service-oriented user interfaces, offering a modular and customizable user experience. Extensive experiments demonstrate an average of 1.59 throughput improvement compared with state-of-the-art baseline. The presented architecture in this work provides actionable insights for next-generation RL training system designs.
This paper has not been read by Pith yet.
Forward citations
Cited by 22 Pith papers
-
Harnessing Routing Foresight for Micro-step-level MoE load balancing in RL Post-training
ForeMoE uses routing foresight from the rollout stage to enable micro-step load balancing in MoE RL post-training via a hierarchical planner and transfer engine, claiming up to 1.45x speedup on 64 GPUs.
-
AstraFlow: Dataflow-Oriented Reinforcement Learning for Agentic LLMs
AstraFlow decouples RL components into autonomous dataflow services to natively support multi-policy agentic LLM training, elastic scaling, and cross-region execution with 2.7x speedup on math, code, search, and Agent...
-
ReCrit: Transition-Aware Reinforcement Learning for Scientific Critic Reasoning
ReCrit frames critic interaction as a correctness-transition problem and uses quadrant-based RL rewards to improve LLM performance on scientific reasoning benchmarks by rewarding corrections and robustness while penal...
-
HetRL: Efficient Reinforcement Learning for LLMs in Heterogeneous Environments
HetRL delivers up to 9.17x higher throughput for LLM RL training on heterogeneous GPUs by using hybrid and ILP-based schedulers to solve a joint optimization problem over computation and data dependencies.
-
RolloutPipe: Overlapping Pipelined Rollout and Training in Disaggregated On-Policy LLM Reinforcement Learning
RolloutPipe uses complete-group pipelining and frontier-group dispatch to overlap rollout and training in disaggregated on-policy RLVR while preserving correctness, cutting end-to-end time by 30.7-42.3% versus Slime.
-
AsyncWebRL: Efficient Multi-Step RL for Visual Web Agents
AsyncWebRL reports up to 2.9x training speedup and new SOTA on WebGym OOD split via async overlap plus constant normalizer in GRPO, with largest gains on harder tasks.
-
Rollout-Level Advantage-Prioritized Experience Replay for GRPO
Rollout-level advantage-prioritized experience replay for GRPO recycles high-advantage individual rollouts with age eviction and fresh-anchored batches to outperform standard GRPO on math benchmarks, with gains increa...
-
How Off-Policy Can GRPO Be? Mu-GRPO for Efficient LLM Reinforcement Learning
Mu-GRPO enables substantially more off-policy GRPO training for LLMs via relaxed clipping and negative-advantage veto in large staged batches, matching standard GRPO performance at ~2x training speed.
-
MinT: Managed Infrastructure for Training and Serving Millions of LLMs
MinT enables efficient management of million-scale LoRA-adapted LLM policies over shared 1T-parameter base models by moving only small adapters through training and serving pipelines.
-
ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
ROSE is a system for cooperative elasticity that co-locates serving and rollout models on shared GPUs, delivering 1.3-3.3x higher end-to-end throughput than fixed-resource baselines while preserving serving SLOs.
-
ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RL
ROSE delivers 1.2-3.3x higher end-to-end throughput for agentic RL by safely co-using underutilized serving GPUs for rollouts while meeting serving SLOs.
-
DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Training
DORA's multi-version streaming rollout enables 2-3x higher throughput in asynchronous RL for LLMs while preserving convergence by maintaining policy consistency, data integrity, and bounded staleness.
-
Relax: An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale
Relax is a new RL training engine with omni-native design and async execution that delivers up to 2x speedups over baselines like veRL while converging to equivalent reward levels on Qwen3 models.
-
Seer: Online Context Learning for Fast Synchronous LLM Reinforcement Learning
Seer improves synchronous LLM RL rollout throughput by up to 2.04x and reduces long-tail latency by 72-94% via divided rollout, context-aware scheduling, and adaptive grouped speculative decoding based on prompt simil...
-
RLBoost: Harvesting Preemptible Resources for Cost-Efficient Reinforcement Learning on LLMs
RLBoost harvests preemptible GPUs for RL rollout via a hybrid architecture with adaptive offload, pull-based transfer, and token-level migration, delivering 1.51x-1.97x throughput and 28-49% better cost efficiency tha...
-
DARTS: Distribution-Aware Active Rollout Trajectory Shaping for Accelerating LLM Reinforcement Learning
DARTS accelerates LLM RL training up to 1.77x by distribution-aware trajectory sampling and adaptive redundancy allocation that shapes rollouts toward conciseness without performance loss.
-
MinT: Managed Infrastructure for Training and Serving Millions of LLMs
MinT is a system for managing million-scale LoRA adapter catalogs on shared 1T-parameter base models, with reported efficiency gains in adapter movement, multi-policy training, and catalog addressability.
-
Next-Generation Agentic Reinforcement Learning Systems Enable Self-Evolving Agents
Current agentic RL systems lack three key components needed for self-evolving agents at scale, requiring new co-designed architectures such as AReaL2.0 to enable policy updates from deployed workloads.
-
Xiaomi-GUI-0 Technical Report
Xiaomi-GUI-0 reports 72.0% success on an in-house real-mobile benchmark and 78.9% on AndroidWorld after training a GUI agent in a real-device closed loop with an error-driven data flywheel and three-stage RL pipeline.
-
Xiaomi-GUI-0 Technical Report
Xiaomi-GUI-0 reports 72.0% success on RealMobile and 78.9% on AndroidWorld via real-device closed-loop training with multi-source data and three-stage RL pipeline.
-
Libra: Efficient Resource Management for Agentic RL Post-Training
Libra optimizes GPU allocation across rollout and training in agentic RL via an elastic hybrid pool and C-MLFQ scheduler based on tool-return causal signals, claiming up to 3.0x throughput and 2.5x faster reward conve...
-
Next-Generation Agentic Reinforcement Learning Systems Enable Self-Evolving Agents
Enterprise self-evolving agents require new agentic RL systems built around standardized trajectory data protocols, workload-to-learning data proxies, and automatic policy evolution control planes.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.