JoyNexus：服務導向的多租戶後訓練於VLA模型

摘要

視覺-語言-動作（VLA）模型的後訓練因模擬器、機器人形態及任務目標的多樣性而至關重要。現有的計算服務——無論是以直接加速器租賃還是批次工作負載提交的形式——通常將一組獨占的GPU和CPU資源分配給單一租戶。此模式雖最大化客戶靈活性，卻迫使使用者承擔基礎設施適配的負擔，而固定的卡時計費模型使得短期或突發性工作負載對租戶而言成本高昂，對服務提供者則效率低落。為應對這些挑戰，我們提出JoyNexus，一個用於多租戶VLA有監督微調、強化學習與評估的統一服務。JoyNexus將訓練模型服務、推理模型服務與環境服務解耦，每項服務均透過API存取，並由駐留的共享基礎模型搭配租戶專屬插槽支援。租戶可直接調用高階語義API進行訓練、滾動執行與評估，或利用低階API及其指定的端點組合自定義演算法。多個租戶並行提交工作負載，其動作模組、優化器、滾動記錄與策略版本保持隔離，服務則由全域訓練佇列與推理佇列調度。為進一步提升多租戶訓練效率，JoyNexus針對共享相容模型前綴的異質VLA資料模式引入群組批次處理，使分組樣本能通過單一共享主幹前向傳播。最後，我們透過工作負載模擬與真實具身場景中的群組批次處理管線評估JoyNexus。結果顯示，相較於隔離的單租戶執行，JoyNexus經由跨租戶調度共享資源，降低了總計GPU時間並提升了服務利用率。

English

The post-training of Vision-Language-Action (VLA) models is essential due to the diversity of simulators, robot embodiments, and task objectives. Existing compute services, whether offered as direct accelerator rental or batch-workload submission, typically allocate an exclusive set of GPU and CPU resources to a single tenant. While this paradigm maximizes client flexibility, it burdens users with infrastructure adaptation, and the fixed card-hour accounting model renders short or bursty workloads both expensive for tenants and inefficient for the service provider. To address these challenges, we present JoyNexus, a unified service for multi-tenant VLA supervised fine-tuning, reinforcement learning, and evaluation. JoyNexus decouples the Training Model Service, Inference Model Service, and Environment Service, each accessed through APIs and backed by resident shared base models with tenant-specific slots. Tenants can directly invoke high-level semantic APIs for training, rollout, and evaluation, or compose custom algorithms using lower-level APIs and their assigned endpoints. Multiple tenants submit workloads concurrently; their action modules, optimizers, rollout records, and policy versions remain isolated, and the service is scheduled by the global Training Queue and Inference Queue. To further improve multi-tenant training efficiency, JoyNexus introduces group batching for heterogeneous VLA data schemas that share a compatible model-facing prefix, enabling a single shared backbone forward pass over grouped samples. Finally, we evaluate JoyNexus through workload simulation and a group-batching pipeline in a realistic embodied scenario. Results show that, compared with isolated single-tenant execution, JoyNexus reduces aggregate GPU time and improves service utilization via cross-tenant scheduling on shared resources.