Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

VisionZip: Longer is Better but Not Necessary in Vision Language Models

About

Recent advancements in vision-language models have enhanced performance by increasing the length of visual tokens, making them much longer than text tokens and significantly raising computational costs. However, we observe that the visual tokens generated by popular vision encoders, such as CLIP and SigLIP, contain significant redundancy. To address this, we introduce VisionZip, a simple yet effective method that selects a set of informative tokens for input to the language model, reducing visual token redundancy and improving efficiency while maintaining model performance. The proposed VisionZip can be widely applied to image and video understanding tasks and is well-suited for multi-turn dialogues in real-world scenarios, where previous methods tend to underperform. Experimental results show that VisionZip outperforms the previous state-of-the-art method by at least 5% performance gains across nearly all settings. Moreover, our method significantly enhances model inference speed, improving the prefilling time by 8x and enabling the LLaVA-Next 13B model to infer faster than the LLaVA-Next 7B model while achieving better results. Furthermore, we analyze the causes of this redundancy and encourage the community to focus on extracting better visual features rather than merely increasing token length. Our code is available at https://github.com/dvlab-research/VisionZip .

Senqiao Yang, Yukang Chen, Zhuotao Tian, Chengyao Wang, Jingyao Li, Bei Yu, Jiaya Jia• 2024

Related benchmarks

TaskDatasetResultRank
Object Hallucination EvaluationPOPE
Accuracy88.48
2019
Visual Question AnsweringVizWiz
Accuracy98.75
1820
Visual Question AnsweringTextVQA
Accuracy62
1453
Visual Question AnsweringVQA v2
Accuracy79.7
1429
Visual Question AnsweringGQA
Accuracy62.4
1425
Automatic Speech RecognitionLibriSpeech clean (test)
WER5.93
1207
Automatic Speech RecognitionLibriSpeech (test-other)
WER7.71
1206
Text-based Visual Question AnsweringTextVQA
Accuracy71.7
962
Multimodal UnderstandingMMBench
Accuracy85.1
847
Science Question AnsweringScienceQA
Accuracy95.07
791
Showing 10 of 499 rows
...

Other info

Code

Follow for update