Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912369147379712 |
|---|---|
| author | Zhang, Qizhe Cheng, Aosong Lu, Ming Zhang, Renrui Zhuo, Zhiyong Cao, Jiajun Guo, Shaobo She, Qi Zhang, Shanghang |
| author_facet | Zhang, Qizhe Cheng, Aosong Lu, Ming Zhang, Renrui Zhuo, Zhiyong Cao, Jiajun Guo, Shaobo She, Qi Zhang, Shanghang |
| contents | Large vision-language models (LVLMs) generally contain significantly more visual tokens than their textual counterparts, resulting in a considerable computational burden. Recent efforts have been made to tackle this issue by pruning visual tokens early within the language model. Most existing works use attention scores between text and visual tokens to assess the importance of visual tokens. However, in this study, we first analyze the text-visual attention in the language model and find that this score is not an ideal indicator for token pruning. Based on the analysis, We propose VisPruner, a plug-and-play method that utilizes visual cues for more effective token pruning in LVLMs. Specifically, we first use visual attention to select a limited number of significant tokens. Then, we remove duplicate tokens from the remaining ones based on their similarity. By retaining diverse tokens alongside the initially selected important tokens, we maximally preserve the visual information of the input image. Experimental results demonstrate that our VisPruner sustains strong performance across various VLM architectures and reduction ratios, significantly outperforming existing methods based on text-visual attention. Notably, without any training, VisPruner can reduce the FLOPs of LLaVA-1.5-7B by 91% and inference latency by 75%, while maintaining comparable performance. Our code is available at https://github.com/Theia-4869/VisPruner. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_01818 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs Zhang, Qizhe Cheng, Aosong Lu, Ming Zhang, Renrui Zhuo, Zhiyong Cao, Jiajun Guo, Shaobo She, Qi Zhang, Shanghang Computer Vision and Pattern Recognition Artificial Intelligence Large vision-language models (LVLMs) generally contain significantly more visual tokens than their textual counterparts, resulting in a considerable computational burden. Recent efforts have been made to tackle this issue by pruning visual tokens early within the language model. Most existing works use attention scores between text and visual tokens to assess the importance of visual tokens. However, in this study, we first analyze the text-visual attention in the language model and find that this score is not an ideal indicator for token pruning. Based on the analysis, We propose VisPruner, a plug-and-play method that utilizes visual cues for more effective token pruning in LVLMs. Specifically, we first use visual attention to select a limited number of significant tokens. Then, we remove duplicate tokens from the remaining ones based on their similarity. By retaining diverse tokens alongside the initially selected important tokens, we maximally preserve the visual information of the input image. Experimental results demonstrate that our VisPruner sustains strong performance across various VLM architectures and reduction ratios, significantly outperforming existing methods based on text-visual attention. Notably, without any training, VisPruner can reduce the FLOPs of LLaVA-1.5-7B by 91% and inference latency by 75%, while maintaining comparable performance. Our code is available at https://github.com/Theia-4869/VisPruner. |
| title | Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2412.01818 |