WeMMU: Enhanced Bridging of Vision-Language Models and Diffusion Models via Noisy Query Tokens
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908687243673600 |
|---|---|
| author | Yang, Jian Yin, Dacheng He, Xiaoxuan Li, Yong Rao, Fengyun Lyu, Jing Zhai, Wei Cao, Yang Zha, Zheng-Jun |
| author_facet | Yang, Jian Yin, Dacheng He, Xiaoxuan Li, Yong Rao, Fengyun Lyu, Jing Zhai, Wei Cao, Yang Zha, Zheng-Jun |
| contents | Recent progress in multimodal large language models (MLLMs) has highlighted the challenge of efficiently bridging pre-trained Vision-Language Models (VLMs) with Diffusion Models. While methods using a fixed number of learnable query tokens offer computational efficiency, they suffer from task generalization collapse, failing to adapt to new tasks that are distant from their pre-training tasks. To overcome this, we propose Noisy Query Tokens, which learn a distributed representation space between the VLM and Diffusion Model via end-to-end optimization, enhancing continual learning. Additionally, we introduce a VAE branch with linear projection to recover fine-grained image details. Experimental results confirm our approach mitigates generalization collapse and enables stable continual learning across diverse tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_02536 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | WeMMU: Enhanced Bridging of Vision-Language Models and Diffusion Models via Noisy Query Tokens Yang, Jian Yin, Dacheng He, Xiaoxuan Li, Yong Rao, Fengyun Lyu, Jing Zhai, Wei Cao, Yang Zha, Zheng-Jun Computer Vision and Pattern Recognition Recent progress in multimodal large language models (MLLMs) has highlighted the challenge of efficiently bridging pre-trained Vision-Language Models (VLMs) with Diffusion Models. While methods using a fixed number of learnable query tokens offer computational efficiency, they suffer from task generalization collapse, failing to adapt to new tasks that are distant from their pre-training tasks. To overcome this, we propose Noisy Query Tokens, which learn a distributed representation space between the VLM and Diffusion Model via end-to-end optimization, enhancing continual learning. Additionally, we introduce a VAE branch with linear projection to recover fine-grained image details. Experimental results confirm our approach mitigates generalization collapse and enables stable continual learning across diverse tasks. |
| title | WeMMU: Enhanced Bridging of Vision-Language Models and Diffusion Models via Noisy Query Tokens |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.02536 |