LightBagel：轻量级双融合框架，实现统一的多模态理解与生成

摘要

近期，统一多模态模型在能力与通用性方面展现出显著提升，但主流系统仍多需从头训练且消耗大量算力资源。本文证明，通过策略性融合专精于生成或理解任务的公开模型，能以更高效率获得具有竞争力的性能。我们的核心设计是在保留原始模块的同时，于整个网络中交错插入多模态自注意力模块。这种双重融合机制具有双重优势：（1）在充分保留基础模型原有优势的前提下实现高效的多模态融合；（2）促使理解编码器的高级语义表征与生成编码器的低级空间信号产生协同融合。该方法仅需约350亿标记的训练量，便在多项基准测试中取得优异成果：组合式文生图任务GenEval得分0.91，复杂文生图任务DPG-Bench得分82.16，图像编辑任务GEditBench与ImgEdit-Bench分别获得6.06和3.77分。我们完整开源代码、模型权重及数据集，以支持统一多模态建模的未来研究。

English

Unified multimodal models have recently shown remarkable gains in both capability and versatility, yet most leading systems are still trained from scratch and require substantial computational resources. In this paper, we show that competitive performance can be obtained far more efficiently by strategically fusing publicly available models specialized for either generation or understanding. Our key design is to retain the original blocks while additionally interleaving multimodal self-attention blocks throughout the networks. This double fusion mechanism (1) effectively enables rich multi-modal fusion while largely preserving the original strengths of the base models, and (2) catalyzes synergistic fusion of high-level semantic representations from the understanding encoder with low-level spatial signals from the generation encoder. By training with only ~ 35B tokens, this approach achieves strong results across multiple benchmarks: 0.91 on GenEval for compositional text-to-image generation, 82.16 on DPG-Bench for complex text-to-image generation, 6.06 on GEditBench, and 3.77 on ImgEdit-Bench for image editing. By fully releasing the entire suite of code, model weights, and datasets, we hope to support future research on unified multimodal modeling.

LightBagel：轻量级双融合框架，实现统一的多模态理解与生成

LightBagel: A Light-weighted, Double Fusion Framework for Unified Multimodal Understanding and Generation

摘要

Support