aiqbench analysis
JoyNexus addresses the inefficiency of current compute services for Vision-Language-Action models by introducing a multi-tenant architecture with shared backbones and isolated tenant slots. This approach decouples training, inference, and environment services to optimize resource utilization for diverse simulators and robot embodiments while maintaining data isolation through global scheduling queues. The system further enhances efficiency via group batching strategies that handle heterogeneous VLA data schemas without compromising model performance.
JoyNexus: Service-Oriented Multi-Tenant Post-Training for VLA Models
arXiv:2607.16074v1 Announce Type: cross Abstract: 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.
Abstract: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.
Distributed, Parallel, and Cluster Computing (cs.DC); Artificial Intelligence (cs.AI); Software Engineering (cs.SE)
arXiv:2607.16074 [cs.DC]
(or arXiv:2607.16074v1 [cs.DC] for this version)
https://doi.org/10.48550/arXiv.2607.16074 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 272 words
Related coverage
News
Quantum Computing Report · Jul 25, 2026
Spur: Modern GPU Job Scheduling for HPC and AI Workloads
AMD · Jul 22, 2026
Share GPU clusters across teams with isolation and fairness using Amazon SageMaker HyperPod
AWS · Oct 8, 2026
Building Blocks for Foundation Model Training and Inference on AWS
Hugging Face · May 11, 2026