Spider: Any-to-Many Multimodal LLM

TL;DR

Spider is a novel framework that enables large language models to generate arbitrary combinations of multiple modalities simultaneously, significantly advancing multimodal interaction capabilities.

Contribution

Introduces Spider, an efficient Any-to-Many Modalities Generation framework with new components and a novel dataset, enabling flexible multimodal content generation beyond pairwise modalities.

Findings

Successfully generates arbitrary modality combinations 'Text + Xs'

Creates the first X-to-Xs many-modal dataset

Enhances multimodal interaction and future research potential

Abstract

Multimodal LLMs (MLLMs) have emerged as an extension of Large Language Models (LLMs), enabling the integration of various modalities. However, Any-to-Any MLLMs are limited to generating pairwise modalities 'Text + X' within a single response, such as Text + {Image or Audio or Video}. To address this limitation, we introduce Spider, a novel efficient Any-to-Many Modalities Generation (AMMG) framework, which can generate an arbitrary combination of modalities 'Text + Xs', such as Text + {Image and Audio and Video}. To achieve efficient AMMG, our Spider integrates three core components: a Base Model for basic X-to-X (i.e., Any-to-Any) modality processing, an Any-to-Many Instruction Template designed for producing Xs signal prompts, and a novel Efficient Decoders-Controller for controlling multimodal Decoders to generate Xs (many-modal) contents. To train Spider, we constructed a novel Text-formatted Many-Modal (TMM) dataset, which facilitates learning the X-to-Xs (i.e., Any-to-Many) capability necessary for AMMG. Ultimately, the well-trained Spider generates a pseudo X-to-Xs dataset, the first-ever X-to-Xs many-modal dataset, enhancing the potential for AMMG tasks in future research. Overall, this work not only pushes the boundary of multimodal interaction but also provides rich data support for advancing the field. Code: https://github.com/Layjins/Spider