Taiyi | Lab Index

First open-source Chinese Stable Diffusion model family. Trained on 20M filtered Chinese image-text pairs from Noah-Wukong and Zero datasets. Taiyi-Diffusion-XL extends to bilingual (Chinese-English) generation with 3.5B parameters via continuous pre-training on SDXL.

Paper (Taiyi-Diffusion-XL, arXiv)HuggingFace (Chinese SD)HuggingFace (XL)

Outputs 3

Taiyi-Stable-Diffusion-1B-Chinese

model

First open-source Chinese Stable Diffusion model trained on 20M filtered Chinese image-text pairs.

HuggingFace

Taiyi-Diffusion-XL

model 2024-01-26

Bilingual Chinese-English text-to-image model based on SDXL with expanded vocabulary and vision-language model enhanced prompts.

HuggingFace

Taiyi-Diffusion-XL: Advancing Bilingual Text-to-Image Generation with Large Vision-Language Model Support

paper 2024-01-26

Extends CLIP and SDXL for bilingual text-to-image generation through vocabulary expansion and bilingual continuous pre-training.

Paper (arXiv)

arXiv: 2401.14688

visiongenerationnlpopen-source