Most of the images in the last thread got nuked for being too realistic (understandable). So I'll do what I was thinking about doing anyways and host images on mega, then use a "converted to drawing" version for the previews here. Just know the actual images in the mega folder are in a realistic style.

The mega link will be https://mega.nz/folder/YzRg2JjD#qcvfmIAnU-OUqsjRnpsjxg

The password for every zip file in there will just be 'gurochan' (no quotes). it's only there to stop mega from sticking their nose in my shit. each zip folder contains a handful of images as well as the original prompt.

Someone asked on the last thread - here's the link to the models:
https://huggingface.co/Comfy-Org/MOSS-Image
(i use moss_image_sft_bf16.safetensors, do the int8 version if you have a smaller gpu)
i am using them with the default "Moss-Image Base: Layout-Guided Text to Image" template available in comfyui. you can also drag the images from the mega into comfyui to load the workflow.

this preview image is one of 75 from the Fantasy_Archer folder in the mega.