If these people have existing Keras code they want to integrate or they are interested in developing it further in Keras, then it shouldn't require any insider knowledge to create a Keras version of a small but popular open-source project like this. I am very sure we'd get a PyTorch version made by outsiders quickly if Stable Diffusion was originally released in Keras/TF.
From what I can tell this Keras version was just released (the date on the post is Sep. 25) and the first author listed is the creator of Keras. Is this incorrect? I am not familiar with Divam Gupta and I would consider outsiders to be people not paid by Google.
I didn't say that it casts a negative light on Keras. Just on its popularity among outsiders. There are thousands of great libraries out there that are much less popular than Keras or PyTorch. And BTW, JAX is a useful Google-created framework that's growing in popularity among researchers and pushed PyTorch to improve (functorch), so I have nothing against Google projects.
So why are Keras creators listed as authors on this post and why is it on Keras' official site? Compare this to hundreds of PyTorch SD forks that have been thrown up on GitHub.
The OP was wondering whether additional enhancements will also be ported and that's what I responding to. It's simply much less likely that a new paper will get a Keras implementation than a PyTorch implementation.
> So why are Keras creators listed as authors on this post and why is it on Keras' official site?
Because when Keras creators learned about the port done by what you define as an outsider they thought that it was cool - and that it would be nice to make it part of the KerasCV toolbox.
What answer did you expect?
You understand that it was "an outsider" who ported Stable Diffusion to Keras, right?
Timeline:
Aug 22, 2022 [day 1 of the SD era] Public release of stable diffusion
Sep 17, 2022 [day 27 of the SD era] Keras outsider Divam Gupta announces a Keras port
"Stable Diffusion implemented using @Tensorflow and #Keras. [...] Thanks @fchollet and team for building this amazing framework which makes it easy to implement a model like Stable Diffusion."
Sep 19, 2022 [day 29 of the SD era, 2 days after port announcement] François Chollet publishes a Twitter thread about the port (and his own improvements on a fork - two dozen commits dated Sep 18-21 will be pushed upstream)
"Huge thanks to @divamgupta for creating this port! This is top-quality work that will benefit everyone doing creative AI. I'm always amazed by the velocity of the open-source community"
Sep 25, 2022 [day 35 of the SD era, 8 days after port] Divam Gupta talks about the upcoming KerasCV integration
"Last week I implemented Stable Diffusion using Keras / Tensorflow. Now its almost integrated in KerasCV thanks to @fchollet and team. This is the power of open source collaboration."
Sep 27, 2022 [day 37 of the SD era, 10 days after port announcement] Official release of the Stable Diffusion implementation in KerasCV - and TFA (dated Sep 25)
"Stable Diffusion is now available directly in KerasCV! [...] Many thanks to all those who made this implementation possible, in particular @divamgupta @luke_wood_ml and of course the creators of the original Stable Diffusion models!"
