vTower-Sanctuary is a 5050 weighted sum merge between <a target="_blank" rel="ugc" href="https://civarchive.com/models/835578?modelVersionId=1029914">PaSanctuary 2.5</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/920709?modelVersionId=1029914">Cat Tower 1.0,</a> please check them out, maybe give them a thumbs up too if you like this merge :)The merge has a clean default style that changes a bit to match the subject matter, and it adheres pretty well to artist styles as most illustrious models do. The usual quality tags work and can affect the output a lot, so play around with them to see how they work and what puts out gens you like. Pretty decent range, not just a girl generator (though that way inclined unless otherwise prompted) and understands your intent pretty well.Recommended settings in the example images, just what I personally found to work well, feel free to play with these parameters too.v3? yeah, there -&gt;v2.0 ? check the description -&gt;I liked two illustrious checkpoints shared here, <a target="_blank" rel="ugc" href="https://civarchive.com/models/813634/silvermoonmix01-illustrious?modelVersionId=954246">SilvermoonMix01-Illustrious 1.1</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/838773/illustriouspencil-xl?modelVersionId=951879">illustrious_pencil-XL 1.1,</a> but had some minor gripes with the default styles of both, so I tried a very simple 50-50 merge between them and was pleasantly surprised by the result. Seems to have gotten rid of the lighting quirks of pencil and face quirks of silvermoon, though both of these might just have been skill issues on my part. I decided to share, since it's possible someone is looking for the same results as me. I've done very little testing overall, but this should still retain the prompt comprehension and artist and character knowledge of original Illustrious.I can recommend the following settings, but feel free to experiment and follow your heart:28 steps 4 cfg euler a - align your stepspositive: masterpiece, best quality,negative: bad anatomy, bad hands, error, extra digits, fewer digits, censor, sunlight, brightIllustrious seems like a good step forward for local 'illustration' generation and I'm looking forward to many more merges and finetunes in the coming months.

Simple Illustrious merge

simpleIllustrious_valt.safetensors

vAlt

vTower-Sanctuary

v3-way

v2.0

v1.0

v2.0 ? check the description -&gt;I liked two illustrious checkpoints shared here, <a target="_blank" rel="ugc" href="https://civitai.com/models/813634/silvermoonmix01-illustrious?modelVersionId=954246">SilvermoonMix01-Illustrious 1.1</a> and <a target="_blank" rel="ugc" href="https://civitai.com/models/838773/illustriouspencil-xl?modelVersionId=951879">illustrious_pencil-XL 1.1,</a> but had some minor gripes with the default styles of both, so I tried a very simple 50-50 merge between them and was pleasantly surprised by the result. Seems to have gotten rid of the lighting quirks of pencil and face quirks of silvermoon, though both of these might just have been skill issues on my part. I decided to share, since it's possible someone is looking for the same results as me. I've done very little testing overall, but this should still retain the prompt comprehension and artist and character knowledge of original Illustrious.I can recommend the following settings, but feel free to experiment and follow your heart:28 steps 4 cfg euler a - align your stepspositive: masterpiece, best quality,negative: bad anatomy, bad hands, error, extra digits, fewer digits, censor, sunlight, brightIllustrious seems like a good step forward for local 'illustration' generation and I'm looking forward to many more merges and finetunes in the coming months.

Files with SHA256: a74ca2858b908ce106a8b11b4eed7d792ac6e6b316874f61959d618c393a27ff