FreqCache: Accelerating Embodied VLN Models with Adaptive Frequency-Guided Token Caching
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914510485323776 |
|---|---|
| author | Zheng, Zihao Zhou, Xingyue Mao, Zhihao Sun, Songyu Zhang, Lingyue Ao, Yulong Feng, Yupu Zhang, Qiongqiong Lin, Yonghua Chen, Xiang |
| author_facet | Zheng, Zihao Zhou, Xingyue Mao, Zhihao Sun, Songyu Zhang, Lingyue Ao, Yulong Feng, Yupu Zhang, Qiongqiong Lin, Yonghua Chen, Xiang |
| contents | Vision-Language-Navigation (VLN) models exhibit excellent navigation accuracy but incur high computational overhead. Token caching has emerged as a promising training-free strategy to reduce this cost by reusing token computation results; however, existing token caching approaches rely on visual domain methods for cacheable token selection, leading to challenges when adapted to VLN models. 1) Visual domain methods become invalid when there is viewpoint migration. 2) Visual domain methods neglect critical edge information without the aid of additional algorithms. 3) Visual domain methods overlook the temporal variation of scenarios and lack adjustability in cache budgets. In this paper, we develop detailed analyses and find that the impacts of these challenges exhibit invariance and analyzability in the frequency domain. Based on these, we propose a frequency-guided token caching framework, called FreqCache. Utilizing the inherent properties of the frequency domain, FreqCache achieves optimal token cache establishment, refreshment, and adaptive adjustment. Experiments show that FreqCache achieves 1.59x speedup with ignorable overhead, showing the effect of integrating frequency domain methods in VLN token caching. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_24391 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | FreqCache: Accelerating Embodied VLN Models with Adaptive Frequency-Guided Token Caching Zheng, Zihao Zhou, Xingyue Mao, Zhihao Sun, Songyu Zhang, Lingyue Ao, Yulong Feng, Yupu Zhang, Qiongqiong Lin, Yonghua Chen, Xiang Robotics Vision-Language-Navigation (VLN) models exhibit excellent navigation accuracy but incur high computational overhead. Token caching has emerged as a promising training-free strategy to reduce this cost by reusing token computation results; however, existing token caching approaches rely on visual domain methods for cacheable token selection, leading to challenges when adapted to VLN models. 1) Visual domain methods become invalid when there is viewpoint migration. 2) Visual domain methods neglect critical edge information without the aid of additional algorithms. 3) Visual domain methods overlook the temporal variation of scenarios and lack adjustability in cache budgets. In this paper, we develop detailed analyses and find that the impacts of these challenges exhibit invariance and analyzability in the frequency domain. Based on these, we propose a frequency-guided token caching framework, called FreqCache. Utilizing the inherent properties of the frequency domain, FreqCache achieves optimal token cache establishment, refreshment, and adaptive adjustment. Experiments show that FreqCache achieves 1.59x speedup with ignorable overhead, showing the effect of integrating frequency domain methods in VLN token caching. |
| title | FreqCache: Accelerating Embodied VLN Models with Adaptive Frequency-Guided Token Caching |
| topic | Robotics |
| url | https://arxiv.org/abs/2604.24391 |