love it! I'll take a look ;)

My Poor Hardrive! xD
Going to bookmark your model for now. 9 Gig is a bit big. but looks good

same with other comment, I do want to come back to this because it looks so nice ..but 8 gig is abit much.

Yes, it looks as a great model but, if possible, can you add the smaller 2gb pruned-fp16 version ?

I'm very intrigued, size maybe means higher quality, so no problems for me. will be interesting to see what comes out of this one, thx!

Is the VAE baked in or do u have a recommendation? Any optimal dimensions?

Suggestion: Run it through A1111 Model toolkit, see if it needs junk and precision tuning. IT may not need it, but 9 gb means some A1111 instances may not want to load it (I've had struggles loading even Kenshi laetly on GPU instances on colab lol) - Really cute tho!

Pruning does not reduce capacity, pruning removes keys from the model that are NEVER used for generation. Converting precission from fp32 to fp16 reduces the "accuracy", but most people run Stable Diffusion in fp16 mode, so it's irrelevant if the model precission is one or the other