这个算法比playgorund v2.5强太多了

PixArt-alpha / PixArt-sigma

PixArt-Σ: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation

GNU Affero General Public License v3.0

1.47k stars 70 forks source link

Closed rwywenyi closed 3 months ago

rwywenyi commented 3 months ago

Playground v2.5宣传出来的效果很强，使用了一下，还不如PixArt-alpha版本的效果，要知道，PixArt-alpha训练数据要少的多得多

lawrence-cj commented 3 months ago

Released. Will keep updating this repo for a long time. Thanks again for you guys support.

xiexiaoshinick commented 3 months ago

强在那里？我测试了一下，生成图像在美学评分，提示词following和图像中文字生成方面与playgroundv2.5差距还很大。华为论文上也承认这一点，就硬吹么？