WeightBridge: An Efficient Weight Transfer Library for Reinforcement Learning
Not on the map. Why?
Abstract (arXiv)
Weight transfer - the propagation of updated parameters from trainers to rollout generators - is becoming an important performance bottleneck in reinforcement learning (RL) systems for LLMs. The central challenge is supporting the diverse trainer and rollout layouts and synchronization requirements of modern RL workloads without sacrificing efficiency. Existing solutions are efficient under some configurations but perform poorly or lack support under others. We present WeightBridge, a flexible, efficient weight-transfer library designed to deliver high performance across diverse RL configurations. WeightBridge first automatically extracts the correspondence between trainer and rollout weight layouts, then plans and executes redundancy-free and load-balanced weight transfer. It exposes a small, general API while coordinating workers across diverse synchronization modes. Across configurations spanning different models, parallelization layouts, and synchronization modes, WeightBridge reduces average GPU stall time by up to 42× over the state-of-the-art open-source RL framework and achieves high performance in all settings. A coding agent was able to integrate WeightBridge into two different RL frameworks without manual guidance, demonstrating the generality and ease of use of its APIs.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 40 of its 57 references that are in the map. First posted on arXiv 21 Sep 2026.
Builds on
40 of 57 references.
References · 40
- 2017Proximal Policy Optimization Algorithmscited by 14,441
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
- 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
- 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,413
- 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 1,980
- 2019ZeRO: Memory Optimizations Toward Training Trillion Parameter Modelscited by 1,278
- 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
- 2026DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligencecited by 877
- 2021Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LMcited by 621
- 2024DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Servingcited by 595
- 2023FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPUcited by 567
- 2022Efficiently Scaling Transformer Inferencecited by 529
28 more
- 2025Kimi K2: Open Agentic Intelligencecited by 523
- 2022DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scalecited by 403
- 2024Better & Faster Large Language Models via Multi-token Predictioncited by 344
- 2025SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolutioncited by 246
- 2025AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoningcited by 234
- 2024Mooncake: A KVCache-centric Disaggregated Architecture for LLM Servingcited by 193
- 2025Kimi Linear: An Expressive, Efficient Attention Architecturecited by 169
- 2023DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scalescited by 107
- 2024RL on Incorrect Synthetic Data Scales the Efficiency of LLM Math Reasoning by Eight-Foldcited by 99
- 2026Kimi K3: Open Frontier Intelligencecited by 99
- 2025StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generationcited by 56
- 2025Prosperity before Collapse: How Far Can Off-Policy RL Reach with Stale Data on LLMs?cited by 49
- 2024Optimizing RLHF Training for Large Language Models with Stage Fusioncited by 41
- 2024NeMo-Aligner: Scalable Toolkit for Efficient Model Alignmentcited by 40
- 2025AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Trainingcited by 37
- 2025Laminar: A Scalable Asynchronous RL Post-Training Frameworkcited by 35
- 2024RL-GPT: Integrating Reinforcement Learning and Code-as-policycited by 32
- 2023AlpaServe: Statistical Multiplexing with Model Parallelism for Deep Learning Servingcited by 31
- 2025RollPacker: Mitigating Long-Tail Rollouts for Fast, Synchronous RL Post-Trainingcited by 31
- 2022Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learningcited by 20
- 2026Learn Hard Problems During RL with Reference Guided Fine-tuningcited by 11
- 2025An Extensible Software Transport Layer for GPU Networkingcited by 10
- 2025RLBoost: Harvesting Preemptible Resources for Cost-Efficient Reinforcement Learning on LLMscited by 10
- 2025fabric-lib: RDMA Point-to-Point Communication for LLM Systemscited by 8
- 2025Stabilizing Off-Policy Training for Long-Horizon LLM Agent via Turn-Level Importance Sampling and Clipping-Triggered Normalizationcited by 8
- 2026Relax: An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scalecited by 7
- 2025Fast LLM Post-training via Decoupled and Fastest-of-N Speculationcited by 4
- 2026TensorHub: Scalable and Elastic Weight Transfer for LLM RL Trainingcited by 4
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2025Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library
- 2025RollArt: Disaggregated Multi-Task Agentic RL Training at Scale
- 2025BroRL: Scaling Reinforcement Learning via Broadened Exploration
- 2026RolloutPipe: Overlapping Pipelined Rollout and Training in Disaggregated On-Policy LLM Reinforcement Learning
- 2026Asymmetric Prompt Weighting for Reinforcement Learning with Verifiable Rewards
- 2026NeMo-DCR: Bit-Exact Delta-Compressed Refit for Scalable Agentic RL at Trillion-Parameter Scale
- 2026FP8-RL: A Practical and Stable Low-Precision Stack for LLM Reinforcement Learning
- 2026Miles v0.1: Production-Level Post-Training