Ola: Pushing the Frontiers of Omni-Modal Language Model

TL;DR

Ola is an open-source omni-modal language model that achieves competitive performance across image, video, and audio understanding by innovative architecture, data strategies, and a progressive training pipeline, advancing the field of multi-modal AI.

Contribution

The paper introduces Ola, a novel omni-modal language model that significantly improves multi-modal understanding and aligns cross-modal representations more effectively than previous open-source models.

Findings

Ola surpasses existing open omni-modal models across all modalities.

Ola achieves performance comparable to specialized models of similar size.

The proposed training strategy enhances cross-modal alignment and understanding.

Abstract

Recent advances in large language models, particularly following GPT-4o, have sparked increasing interest in developing omni-modal models capable of understanding more modalities. While some open-source alternatives have emerged, there is still a notable lag behind specialized single-modality models in performance. In this paper, we present Ola, an Omni-modal Language model that achieves competitive performance across image, video, and audio understanding compared to specialized counterparts, pushing the frontiers of the omni-modal language model to a large extent. We conduct a comprehensive exploration of architectural design, data curation, and training strategies essential for building a robust omni-modal model. Ola incorporates advanced visual understanding and audio recognition capabilities through several critical and effective improvements over mainstream baselines. Moreover, we rethink inter-modal relationships during omni-modal training, emphasizing cross-modal alignment with video as a central bridge, and propose a progressive training pipeline that begins with the most distinct modalities and gradually moves towards closer modality alignment. Extensive experiments demonstrate that Ola surpasses existing open omni-modal LLMs across all modalities while achieving highly competitive performance compared to state-of-the-art specialized models of similar sizes. We aim to make Ola a fully open omni-modal understanding solution to advance future research in this emerging field. Model weights, code, and data are open-sourced at https://github.com/Ola-Omni/Ola.