NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866908818082889728 |
|---|---|
| author | Chavhan, Ruchika Chadwick, Malcolm Ramos, Alberto Gil Couto Pimentel Morreale, Luca Noroozi, Mehdi Mehrotra, Abhinav |
| author_facet | Chavhan, Ruchika Chadwick, Malcolm Ramos, Alberto Gil Couto Pimentel Morreale, Luca Noroozi, Mehdi Mehrotra, Abhinav |
| contents | While large-scale text-to-image diffusion models continue to improve in visual quality, their increasing scale has widened the gap between state-of-the-art models and on-device solutions. To address this gap, we introduce NanoFLUX, a 2.4B text-to-image flow-matching model distilled from 17B FLUX.1-Schnell using a progressive compression pipeline designed to preserve generation quality. Our contributions include: (1) A model compression strategy driven by pruning redundant components in the diffusion transformer, reducing its size from 12B to 2B; (2) A ResNet-based token downsampling mechanism that reduces latency by allowing intermediate blocks to operate on lower-resolution tokens while preserving high-resolution processing elsewhere; (3) A novel text encoder distillation approach that leverages visual signals from early layers of the denoiser during sampling. Empirically, NanoFLUX generates 512 x 512 images in approximately 2.5 seconds on mobile devices, demonstrating the feasibility of high-quality on-device text-to-image generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_06879 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices Chavhan, Ruchika Chadwick, Malcolm Ramos, Alberto Gil Couto Pimentel Morreale, Luca Noroozi, Mehdi Mehrotra, Abhinav Computer Vision and Pattern Recognition Artificial Intelligence While large-scale text-to-image diffusion models continue to improve in visual quality, their increasing scale has widened the gap between state-of-the-art models and on-device solutions. To address this gap, we introduce NanoFLUX, a 2.4B text-to-image flow-matching model distilled from 17B FLUX.1-Schnell using a progressive compression pipeline designed to preserve generation quality. Our contributions include: (1) A model compression strategy driven by pruning redundant components in the diffusion transformer, reducing its size from 12B to 2B; (2) A ResNet-based token downsampling mechanism that reduces latency by allowing intermediate blocks to operate on lower-resolution tokens while preserving high-resolution processing elsewhere; (3) A novel text encoder distillation approach that leverages visual signals from early layers of the denoiser during sampling. Empirically, NanoFLUX generates 512 x 512 images in approximately 2.5 seconds on mobile devices, demonstrating the feasibility of high-quality on-device text-to-image generation. |
| title | NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2602.06879 |