30 点作者 lnyan大约 3 年前

3 条评论

mdrzn大约 3 年前

Wow, we're getting more and more closer to not ever needing to get a "stock" image again. We'll just generate whatever we need on the fly.

DrSiemer大约 3 年前

Interesting tool, but their video has the worst example of soulless stock music in it I've ever heard

leobg大约 3 年前

Trained on Olivia Frischer , apparently.

Text2Human: Text-Driven Controllable Human Image Generation