Salvato in:
Dettagli Bibliografici
Autori principali: Cai, Wenqi, Zou, Yawen, Li, Guang, Gu, Chunzhi, Zhang, Chao
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2603.07476
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917322461020160
author Cai, Wenqi
Zou, Yawen
Li, Guang
Gu, Chunzhi
Zhang, Chao
author_facet Cai, Wenqi
Zou, Yawen
Li, Guang
Gu, Chunzhi
Zhang, Chao
contents Dataset distillation (DD) aims to synthesize compact training sets that enable models to achieve high accuracy with significantly fewer samples. Recent diffusion-based DD methods commonly introduce semantic guidance through late-stage cross-attention, where textual prompts tend to dominate the generative process. Although this strategy enforces label relevance, it diminishes the contribution of visual latents, resulting in over-corrected samples that mirror prompt patterns rather than reflecting intrinsic visual features. To solve this problem, we introduce an Early Vision-Language Fusion (EVLF) method that aligns textual and visual embeddings at the transition between the encoder and the generative backbone. By incorporating a lightweight cross-attention module at this transition, the early representations simultaneously encode local textures and global semantic directions across the denoising process. Importantly, EVLF is plug-and-play and can be easily integrated into any diffusion-based dataset distillation pipeline with an encoder. It works across different denoiser architectures and sampling schedules without any task-specific modifications. Extensive experiments demonstrate that EVLF generates semantically faithful and visually coherent synthetic data, yielding consistent improvements in downstream classification accuracy across varied settings. Source code is available at https://github.com/wenqi-cai297/earlyfusion-for-dd/.
format Preprint
id arxiv_https___arxiv_org_abs_2603_07476
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle EVLF: Early Vision-Language Fusion for Generative Dataset Distillation
Cai, Wenqi
Zou, Yawen
Li, Guang
Gu, Chunzhi
Zhang, Chao
Computer Vision and Pattern Recognition
Dataset distillation (DD) aims to synthesize compact training sets that enable models to achieve high accuracy with significantly fewer samples. Recent diffusion-based DD methods commonly introduce semantic guidance through late-stage cross-attention, where textual prompts tend to dominate the generative process. Although this strategy enforces label relevance, it diminishes the contribution of visual latents, resulting in over-corrected samples that mirror prompt patterns rather than reflecting intrinsic visual features. To solve this problem, we introduce an Early Vision-Language Fusion (EVLF) method that aligns textual and visual embeddings at the transition between the encoder and the generative backbone. By incorporating a lightweight cross-attention module at this transition, the early representations simultaneously encode local textures and global semantic directions across the denoising process. Importantly, EVLF is plug-and-play and can be easily integrated into any diffusion-based dataset distillation pipeline with an encoder. It works across different denoiser architectures and sampling schedules without any task-specific modifications. Extensive experiments demonstrate that EVLF generates semantically faithful and visually coherent synthetic data, yielding consistent improvements in downstream classification accuracy across varied settings. Source code is available at https://github.com/wenqi-cai297/earlyfusion-for-dd/.
title EVLF: Early Vision-Language Fusion for Generative Dataset Distillation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.07476