Lyra: An Efficient and Speech-Centric Framework for Omni-Cognition

TL;DR

Lyra is a novel multi-modal large language model that integrates speech, vision, and language to achieve efficient, versatile omni-cognition with state-of-the-art performance and reduced resource requirements.

Contribution

Lyra introduces a speech-centric, multi-modal framework utilizing open-source models, a multi-modality LoRA, and a large multi-modal dataset to enhance omni-cognition efficiently.

Findings

Achieves state-of-the-art results on multiple benchmarks.

Handles complex long speech inputs effectively.

Uses fewer resources than comparable models.

Abstract

As Multi-modal Large Language Models (MLLMs) evolve, expanding beyond single-domain capabilities is essential to meet the demands for more versatile and efficient AI. However, previous omni-models have insufficiently explored speech, neglecting its integration with multi-modality. We introduce Lyra, an efficient MLLM that enhances multimodal abilities, including advanced long-speech comprehension, sound understanding, cross-modality efficiency, and seamless speech interaction. To achieve efficiency and speech-centric capabilities, Lyra employs three strategies: (1) leveraging existing open-source large models and a proposed multi-modality LoRA to reduce training costs and data requirements; (2) using a latent multi-modality regularizer and extractor to strengthen the relationship between speech and other modalities, thereby enhancing model performance; and (3) constructing a high-quality, extensive dataset that includes 1.5M multi-modal (language, vision, audio) data samples and 12K long speech samples, enabling Lyra to handle complex long speech inputs and achieve more robust omni-cognition. Compared to other omni-methods, Lyra achieves state-of-the-art performance on various vision-language, vision-speech, and speech-language benchmarks, while also using fewer computational resources and less training data.