NExT-GPT: Any-to-Any Multimodal LLM

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wu, Shengqiong, Fei, Hao, Qu, Leigang, Ji, Wei, Chua, Tat-Seng
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913402929020928
author Wu, Shengqiong
Fei, Hao
Qu, Leigang
Ji, Wei
Chua, Tat-Seng
author_facet Wu, Shengqiong
Fei, Hao
Qu, Leigang
Ji, Wei
Chua, Tat-Seng
contents While recently Multimodal Large Language Models (MM-LLMs) have made exciting strides, they mostly fall prey to the limitation of only input-side multimodal understanding, without the ability to produce content in multiple modalities. As we humans always perceive the world and communicate with people through various modalities, developing any-to-any MM-LLMs capable of accepting and delivering content in any modality becomes essential to human-level AI. To fill the gap, we present an end-to-end general-purpose any-to-any MM-LLM system, NExT-GPT. We connect an LLM with multimodal adaptors and different diffusion decoders, enabling NExT-GPT to perceive inputs and generate outputs in arbitrary combinations of text, images, videos, and audio. By leveraging the existing well-trained highly-performing encoders and decoders, NExT-GPT is tuned with only a small amount of parameter (1%) of certain projection layers, which not only benefits low-cost training and also facilitates convenient expansion to more potential modalities. Moreover, we introduce a modality-switching instruction tuning (MosIT) and manually curate a high-quality dataset for MosIT, based on which NExT-GPT is empowered with complex cross-modal semantic understanding and content generation. Overall, our research showcases the promising possibility of building an AI agent capable of modeling universal modalities, paving the way for more human-like AI research in the community. Project page: https://next-gpt.github.io/
format Preprint
id arxiv_https___arxiv_org_abs_2309_05519
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle NExT-GPT: Any-to-Any Multimodal LLM
Wu, Shengqiong
Fei, Hao
Qu, Leigang
Ji, Wei
Chua, Tat-Seng
Artificial Intelligence
Computation and Language
Machine Learning
While recently Multimodal Large Language Models (MM-LLMs) have made exciting strides, they mostly fall prey to the limitation of only input-side multimodal understanding, without the ability to produce content in multiple modalities. As we humans always perceive the world and communicate with people through various modalities, developing any-to-any MM-LLMs capable of accepting and delivering content in any modality becomes essential to human-level AI. To fill the gap, we present an end-to-end general-purpose any-to-any MM-LLM system, NExT-GPT. We connect an LLM with multimodal adaptors and different diffusion decoders, enabling NExT-GPT to perceive inputs and generate outputs in arbitrary combinations of text, images, videos, and audio. By leveraging the existing well-trained highly-performing encoders and decoders, NExT-GPT is tuned with only a small amount of parameter (1%) of certain projection layers, which not only benefits low-cost training and also facilitates convenient expansion to more potential modalities. Moreover, we introduce a modality-switching instruction tuning (MosIT) and manually curate a high-quality dataset for MosIT, based on which NExT-GPT is empowered with complex cross-modal semantic understanding and content generation. Overall, our research showcases the promising possibility of building an AI agent capable of modeling universal modalities, paving the way for more human-like AI research in the community. Project page: https://next-gpt.github.io/
title NExT-GPT: Any-to-Any Multimodal LLM
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2309.05519