Saved in:
Bibliographic Details
Main Authors: Peng, Tianshuo, Li, Zuchao, Zhang, Lefei, Zhao, Hai, Wang, Ping, Du, Bo
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.07720
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913513273819136
author Peng, Tianshuo
Li, Zuchao
Zhang, Lefei
Zhao, Hai
Wang, Ping
Du, Bo
author_facet Peng, Tianshuo
Li, Zuchao
Zhang, Lefei
Zhao, Hai
Wang, Ping
Du, Bo
contents Large Language Models (LLMs), benefiting from the auto-regressive modelling approach performed on massive unannotated texts corpora, demonstrates powerful perceptual and reasoning capabilities. However, as for extending auto-regressive modelling to multi-modal scenarios to build Large Multi-modal Models (LMMs), there lies a great difficulty that the image information is processed in the LMM as continuous visual embeddings, which cannot obtain discrete supervised labels for classification.In this paper, we successfully perform multi-modal auto-regressive modeling with a unified objective for the first time.Specifically, we propose the concept of visual tokens, which maps the visual features to probability distributions over LLM's vocabulary, providing supervision information for visual modelling.We further explore the distribution of visual features in the semantic space within LMM and the possibility of using text embeddings to represent visual information.Experimental results and ablation studies on 5 VQA tasks and 4 benchmark toolkits validate the powerful performance of our proposed approach.
format Preprint
id arxiv_https___arxiv_org_abs_2403_07720
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Multi-modal Auto-regressive Modeling via Visual Words
Peng, Tianshuo
Li, Zuchao
Zhang, Lefei
Zhao, Hai
Wang, Ping
Du, Bo
Computer Vision and Pattern Recognition
Artificial Intelligence
Large Language Models (LLMs), benefiting from the auto-regressive modelling approach performed on massive unannotated texts corpora, demonstrates powerful perceptual and reasoning capabilities. However, as for extending auto-regressive modelling to multi-modal scenarios to build Large Multi-modal Models (LMMs), there lies a great difficulty that the image information is processed in the LMM as continuous visual embeddings, which cannot obtain discrete supervised labels for classification.In this paper, we successfully perform multi-modal auto-regressive modeling with a unified objective for the first time.Specifically, we propose the concept of visual tokens, which maps the visual features to probability distributions over LLM's vocabulary, providing supervision information for visual modelling.We further explore the distribution of visual features in the semantic space within LMM and the possibility of using text embeddings to represent visual information.Experimental results and ablation studies on 5 VQA tasks and 4 benchmark toolkits validate the powerful performance of our proposed approach.
title Multi-modal Auto-regressive Modeling via Visual Words
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2403.07720