TEAL: Tokenize and Embed ALL for Multi-modal Large Language Models
This addresses the challenge of integrating multiple modalities like image and audio into LLMs for researchers and developers, though it appears incremental as it builds on existing tokenization methods.
The authors tackled the problem of inefficient modeling of multi-modal interactions and non-textual generation in Multi-modal Large Language Models (MM-LLMs) by proposing TEAL, which tokenizes and embeds all modalities into a joint space, enabling frozen LLMs to handle non-textual tasks. Experiments showed substantial improvements in multi-modal understanding and a simple scheme for multi-modal generation.
Despite Multi-modal Large Language Models (MM-LLMs) have made exciting strides recently, they are still struggling to efficiently model the interactions among multi-modal inputs and the generation in non-textual modalities. In this work, we propose TEAL (Tokenize and Embed ALl)}, an approach to treat the input from any modality as a token sequence and learn a joint embedding space for all modalities. Specifically, for the input from any modality, TEAL first discretizes it into a token sequence with the off-the-shelf tokenizer and embeds the token sequence into a joint embedding space with a learnable embedding matrix. MM-LLMs just need to predict the multi-modal tokens autoregressively as the textual LLMs do. Finally, the corresponding de-tokenizer is applied to generate the output in each modality based on the predicted token sequence. With the joint embedding space, TEAL enables the frozen LLMs to perform both understanding and generation tasks involving non-textual modalities, such as image and audio. Thus, the textual LLM can just work as an interface and maintain its high performance in textual understanding and generation. Experiments show that TEAL achieves substantial improvements in multi-modal understanding, and implements a simple scheme for multi-modal generations.