NPU Offloading of a Frozen Visual Encoder for Robot Policy Training

[2608.15002] NPU Offloading of a Frozen Visual Encoder for Robot Policy Training Skip to main content Search Submit Donate Log in Search arXiv Press Enter to search · Advanced search Computer Science > Robotics arXiv:2608.15002 (cs) [Submitted on 15 Aug 2026] Title: NPU Offloading of a Frozen Visual Encoder for Robot Policy Training Authors: Hyojun Yun , Seungjae Won , Hyungpil Moon View a PDF of the paper titled NPU Offloading of a Frozen Visual Encoder for Robot Policy Training, by Hyojun Yun and 2 other authors View PDF HTML (experimental) Abstract: When a robot policy is trained for a new task or dataset, its visual encoder can be frozen and only its action generation module trained, reducing training cost. Freezing removes the encoder's backward pass, but its forward pass must still run at every training step because the input images change, so it keeps consuming GPU compute. We therefore ask whether moving this computation to a low power AI accelerator such as an NPU can reduce total energy despite the added data transfer and longer training time, and how it affects policy performance. We built an asynchronous training pipeline that uses both a GPU and an NPU for the AR-Actor specialist. The frozen visual encoder runs in A8W8 INT8 on a Mobilint Aries2 NPU, while the FP32 action expert is trained on an NVIDIA GeForce RTX 5060 Ti GPU. We compared a GPU-only baseline with four conditions, L1 to L4, which gradually extend NPU offloading from one to four Transformer encoder layers. Each condition was trained for 30,000 steps with three random seeds. We measured GPU board power for the GPU-only condition and combined GPU and NPU board power for the NPU conditions. Energy per sample decreased by 17.1% in L1, which offloaded ResNet18 and the first encoder layer, and by 27.9% in L4, which offloaded ResNet18 and all four encoder layers. In contrast, training time per sample increased by 15.2% in L1 and 37.7% in L4, and peak allocated GPU memory decreased by 19.8 to 20.7%. The 15 resulting policies were each evaluated with the same 300 environment seeds, for a total of 4,500 simulator rollouts. The combined success rate was 93.33% for GPU-only and 91.44 to 92.89% for the NPU conditions. These results show that NPU offloading of a frozen visual encoder can reduce training energy, but it increases training time and lowers policy success rate by 0.44 to 1.89 percentage points compared with GPU-only training. Comments: 6 pages, 4 figures, 4 tables Subjects: Robotics (cs.RO) ; Hardware Architecture (cs.AR); Machine Learning (cs.LG) Cite as: arXiv:2608.15002 [cs.RO] (or arXiv:2608.15002v1 [cs.RO] for this version) https://doi.org/10.48550/arXiv.2608.15002 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Hyojun Yun [ view email ] [v1] Sat, 15 Aug 2026 03:19:09 UTC (500 KB) Full-text links: Access Paper: View a PDF of the paper titled NPU Offloading of a Frozen Visual Encoder for Robot Policy Training, by Hyojun Yun and 2 other authors View PDF HTML (experimental) TeX Source view license Current browse context: cs.RO < prev | next > new | recent | 2026-08 Change to browse by: cs cs.AR cs.LG References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation × loading... Data provided by: Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• NPU Offloading of a Frozen Visual Encoder for Robot Policy Training#612437
• Learning to Fly: Stable Vision-Guided UAV Servoing with Compact Target-Centric Cues and Reinforcement Learning#987000
• Veo-Act: Enhancing VLA Policies with Frontier Video Models#1012179
• HaWMPO: Hallucination-Aware World Model-based Policy Optimization for Generalist Robot Policy#996753
• RLinf-USER: A Unified and Extensible System for Real-World Online Policy Learning in Embodied AI#798763