Hacker News Reader

Not products, but Google Research already published papers about two different, better models:

https://parti.research.google/ https://imagen.research.google/

The models themselves are not public however.

Wow. Those models, particularly Imagen, are of an entirely separate calibre. None of that psychedelic foggy memory swirl that is characteristic of the space. I can see why Google Research is hesitant to release them.

After using all of the different models extensively, Stable Diffusion is currently state-of-the-art.

Images are more artistic and less clip art-like than Dall-E, but also don’t have a house style like Midjourney. It’s stunningly good - and open source.

What’s really cool is that the devs have worked hard to optimise the model, so after being trained on 1000 A100s it’ll run happily on an 8gb graphics card or M2 Mac.

wombo.ai and midjourney