NinA: Normalizing Flows in Action. Training VLA Models with Normalizing Flows
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912645531041792 |
|---|---|
| author | Tarasov, Denis Nikulin, Alexander Zisman, Ilya Klepach, Albina Lyubaykin, Nikita Polubarov, Andrei Derevyagin, Alexander Kurenkov, Vladislav |
| author_facet | Tarasov, Denis Nikulin, Alexander Zisman, Ilya Klepach, Albina Lyubaykin, Nikita Polubarov, Andrei Derevyagin, Alexander Kurenkov, Vladislav |
| contents | Recent advances in Vision-Language-Action (VLA) models have established a two-component architecture, where a pre-trained Vision-Language Model (VLM) encodes visual observations and task descriptions, and an action decoder maps these representations to continuous actions. Diffusion models have been widely adopted as action decoders due to their ability to model complex, multimodal action distributions. However, they require multiple iterative denoising steps at inference time or downstream techniques to speed up sampling, limiting their practicality in real-world settings where high-frequency control is crucial. In this work, we present NinA (Normalizing Flows in Action), a fast and expressive alternative to diffusion-based decoders for VLAs. NinA replaces the diffusion action decoder with a Normalizing Flow (NF) that enables one-shot sampling through an invertible transformation, significantly reducing inference time. We integrate NinA into the FLOWER VLA architecture and fine-tune on the LIBERO benchmark. Our experiments show that NinA matches the performance of its diffusion-based counterpart under the same training regime, while achieving substantially faster inference. These results suggest that NinA offers a promising path toward efficient, high-frequency VLA control without compromising performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_16845 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | NinA: Normalizing Flows in Action. Training VLA Models with Normalizing Flows Tarasov, Denis Nikulin, Alexander Zisman, Ilya Klepach, Albina Lyubaykin, Nikita Polubarov, Andrei Derevyagin, Alexander Kurenkov, Vladislav Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Recent advances in Vision-Language-Action (VLA) models have established a two-component architecture, where a pre-trained Vision-Language Model (VLM) encodes visual observations and task descriptions, and an action decoder maps these representations to continuous actions. Diffusion models have been widely adopted as action decoders due to their ability to model complex, multimodal action distributions. However, they require multiple iterative denoising steps at inference time or downstream techniques to speed up sampling, limiting their practicality in real-world settings where high-frequency control is crucial. In this work, we present NinA (Normalizing Flows in Action), a fast and expressive alternative to diffusion-based decoders for VLAs. NinA replaces the diffusion action decoder with a Normalizing Flow (NF) that enables one-shot sampling through an invertible transformation, significantly reducing inference time. We integrate NinA into the FLOWER VLA architecture and fine-tune on the LIBERO benchmark. Our experiments show that NinA matches the performance of its diffusion-based counterpart under the same training regime, while achieving substantially faster inference. These results suggest that NinA offers a promising path toward efficient, high-frequency VLA control without compromising performance. |
| title | NinA: Normalizing Flows in Action. Training VLA Models with Normalizing Flows |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2508.16845 |