ImageBind-LLM: Multi-modality Instruction Tuning

TL;DR

ImageBind-LLM introduces a multi-modality instruction tuning method for large language models, enabling them to respond to diverse modalities like audio, 3D, and video through a novel embedding alignment and visual cache system.

Contribution

The paper presents a new multi-modality instruction tuning approach for LLMs using ImageBind, allowing responses to multiple modalities with a simple training process and a training-free cache for enhanced performance.

Findings

Model responds effectively to diverse modalities.

Achieves superior multi-modality instruction-following.

Demonstrates high language generation quality across modalities.

Abstract

We present ImageBind-LLM, a multi-modality instruction tuning method of large language models (LLMs) via ImageBind. Existing works mainly focus on language and image instruction tuning, different from which, our ImageBind-LLM can respond to multi-modality conditions, including audio, 3D point clouds, video, and their embedding-space arithmetic by only image-text alignment training. During training, we adopt a learnable bind network to align the embedding space between LLaMA and ImageBind's image encoder. Then, the image features transformed by the bind network are added to word tokens of all layers in LLaMA, which progressively injects visual instructions via an attention-free and zero-initialized gating mechanism. Aided by the joint embedding of ImageBind, the simple image-text training enables our model to exhibit superior multi-modality instruction-following capabilities. During inference, the multi-modality inputs are fed into the corresponding ImageBind encoders, and processed by a proposed visual cache model for further cross-modal embedding enhancement. The training-free cache model retrieves from three million image features extracted by ImageBind, which effectively mitigates the training-inference modality discrepancy. Notably, with our approach, ImageBind-LLM can respond to instructions of diverse modalities and demonstrate significant language generation quality. Code is released at https://github.com/OpenGVLab/LLaMA-Adapter.