The IL version has been released.
The base model was borrowed and trained using illustriousRealismBy_v10VAE.
It is mainly trained on Japanese users.
The files used are almost the same as the SD1.5 version, but the tags and image size have been significantly revised.
The IL version uses trigger words.
The accuracy of generation may not be very good. Please try it a few times to find one that you are satisfied with.
I haven't tried various prompts, so I may make new discoveries.
SD1.5
Weight strength: About 0.75 Sampling steps: 15 No Trigger Use Model: chilloutmix_NiPrunedFp32Fix
This is a live-action opening device.
It mainly works on the upper body.
Works from the front or at a slightly diagonal angle.
My tongue sticks out.