Saved in:
Bibliographic Details
Main Authors: Liu, Zheng, Liang, Hao, Li, Bozhou, Xiong, Wentao, Chen, Chong, He, Conghui, Zhang, Wentao, Cui, Bin
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.20756
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916890737115136
author Liu, Zheng
Liang, Hao
Li, Bozhou
Xiong, Wentao
Chen, Chong
He, Conghui
Zhang, Wentao
Cui, Bin
author_facet Liu, Zheng
Liang, Hao
Li, Bozhou
Xiong, Wentao
Chen, Chong
He, Conghui
Zhang, Wentao
Cui, Bin
contents Vision-Language Models (VLMs) have recently emerged, demonstrating remarkable vision-understanding capabilities. However, training these models requires large-scale datasets, which brings challenges related to efficiency, effectiveness, and quality of web data. In this paper, we introduce SynthVLM, a new data synthesis and curation method for generating image-caption pairs. Unlike traditional methods, where captions are generated from images, SynthVLM utilizes advanced diffusion models and high-quality captions to synthesize and select images from text captions, thereby creating precisely aligned image-text pairs. We further introduce SynthVLM-100K, a high-quality dataset consisting of 100K curated and synthesized image-caption pairs. In both model and human evaluations, SynthVLM-100K outperforms traditional real-world datasets. Leveraging this dataset, we develop a new family of multimodal large language models (MLLMs), SynthVLM-7B and SynthVLM-13B, which achieve state-of-the-art (SOTA) performance on various vision question-answering (VQA) tasks. Notably, our models outperform LLaVA across most metrics with only 18\% pretrain data. Furthermore, SynthVLM-7B and SynthVLM-13B attain SOTA performance on the MMLU benchmark, demonstrating that the high-quality SynthVLM-100K dataset preserves language abilities.
format Preprint
id arxiv_https___arxiv_org_abs_2407_20756
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SynthVLM: Towards High-Quality and Efficient Synthesis of Image-Caption Datasets for Vision-Language Models
Liu, Zheng
Liang, Hao
Li, Bozhou
Xiong, Wentao
Chen, Chong
He, Conghui
Zhang, Wentao
Cui, Bin
Computer Vision and Pattern Recognition
Computation and Language
Vision-Language Models (VLMs) have recently emerged, demonstrating remarkable vision-understanding capabilities. However, training these models requires large-scale datasets, which brings challenges related to efficiency, effectiveness, and quality of web data. In this paper, we introduce SynthVLM, a new data synthesis and curation method for generating image-caption pairs. Unlike traditional methods, where captions are generated from images, SynthVLM utilizes advanced diffusion models and high-quality captions to synthesize and select images from text captions, thereby creating precisely aligned image-text pairs. We further introduce SynthVLM-100K, a high-quality dataset consisting of 100K curated and synthesized image-caption pairs. In both model and human evaluations, SynthVLM-100K outperforms traditional real-world datasets. Leveraging this dataset, we develop a new family of multimodal large language models (MLLMs), SynthVLM-7B and SynthVLM-13B, which achieve state-of-the-art (SOTA) performance on various vision question-answering (VQA) tasks. Notably, our models outperform LLaVA across most metrics with only 18\% pretrain data. Furthermore, SynthVLM-7B and SynthVLM-13B attain SOTA performance on the MMLU benchmark, demonstrating that the high-quality SynthVLM-100K dataset preserves language abilities.
title SynthVLM: Towards High-Quality and Efficient Synthesis of Image-Caption Datasets for Vision-Language Models
topic Computer Vision and Pattern Recognition
Computation and Language
url https://arxiv.org/abs/2407.20756