論文 深掘り Hugging Face 発表: 2026-05-12 HF ↑30

Qwen-Image-VAE-2.0 Technical Report

Qwen-Image-VAE-2.0 Technical Report

著者: Zekai Zhang, Deqing Li, Kuan Cao, Yujia Wu, Chenfei Wu ほか25名

要約

We present Qwen-Image-VAE-2.0, a suite of high-compression Variational Autoencoders (VAEs) that achieve significant advances in both reconstruction fidelity and diffusability. To address the reconstruction bottlenecks of high compression, we adopt an improved architecture featuring Global Skip Conne…

#benchmark#diffusion#alignment#coding

同じカテゴリの記事