If you want to support me, bidding buzz on the weekly <a target="_blank" rel="ugc" href="https://civarchive.com/auctions">auctions page</a> for the model helps me keep it up for on-site generation.<h2 id="recommended-settings-:">Recommended settings :</h2><ul><li>Positive prompt : masterpiece, best quality</li><li>Negative prompt : worst quality, low quality, bad anatomy, watermark, censored</li><li>Sampler : Euler a</li><li>CFG Scale : 5-7</li></ul>I highly recommend using the <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">ADetailer</a> extension for better faces and eyes.For v2, I merged v1 with some NoobAI models, @GarejiSama's <a target="_blank" rel="ugc" href="https://civarchive.com/models/1649312?modelVersionId=1866796">AdAstra</a> and @Manityro's <a target="_blank" rel="ugc" href="https://civarchive.com/models/1732765/vermilion">Vermilion</a>, making v2 much more compatible with NoobAI based LoRAs. The baked in VAE is Anzhc's, which you can get <a target="_blank" rel="ugc" href="https://huggingface.co/Anzhc/Anzhcs-VAEs/blob/main/SDXL%20Anime%20VAE%20Dec-only%20B3.safetensors">here</a>, gives a much sharper look to images, reduced noise, and makes colors less bland overall. LustrousMix v2 generations are more colorful than v1's, with more contrast and sharpness. There is a slight art style change as well, with more mature faces. <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/babdaa02-2197-4a29-9b37-13d2467db4e2/width=525/babdaa02-2197-4a29-9b37-13d2467db4e2.jpeg" /><h3 id="hope-you-enjoy!">Hope you enjoy!</h3>

Hoseki - LustrousMix [IllustriousXL/NoobAI EPS1.1]

hosekiLustrousmix_illustriousV1.safetensors

Illustrious v1

Illust/NoobAI EPS1.1 - v2

Illustrious based merge with a 2D anime style base look. It has a more classical anime aesthetic to it compared to its PonyXL namesake, and a softer look overall. Artist tags work.Recommended settings :<ul><li>Positive prompt : masterpiece, best quality</li><li>Negative prompt : (lowres:1.2), (worst quality:1.4), (low quality:1.4), (bad anatomy:1.4), bad hands, multiple views, comic, jpeg artifacts, patreon logo, patreon username, web address, signature, watermark, text, logo, artist name, censored</li><li>Sampler : Euler a</li><li>CFG Scale : 5-6</li></ul>I highly recommend using the <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">ADetailer</a> extension for better faces and eyes.VAE is baked in.I've included the different models I used in the merge in the suggested resources, check them out!

Hoseki - LustrousMix [IllustriousXL]

Illustrious based merge with a 2D anime style base look. It has a more classical anime aesthetic to it compared to its PonyXL namesake, and a softer look overall. Artist tags work.

Recommended settings :
Positive prompt : masterpiece, best quality

Negative prompt : (lowres:1.2), (worst quality:1.4), (low quality:1.4), (bad anatomy:1.4), bad hands, multiple views, comic, jpeg artifacts, patreon logo, patreon username, web address, signature, watermark, text, logo, artist name, censored

Sampler : Euler a

CFG Scale : 5-6

I highly recommend using the ADetailer extension for better faces and eyes.
VAE is baked in.
I've included the different models I used in the merge in the suggested resources, check them out!

If you want to support me, bidding buzz on the weekly <a target="_blank" rel="ugc" href="https://civitai.com/auctions">auctions page</a> for the model helps me keep it up for on-site generation.<h2 id="recommended-settings-:">Recommended settings :</h2><ul><li>Positive prompt : masterpiece, best quality</li><li>Negative prompt : worst quality, low quality, bad anatomy, watermark, censored</li><li>Sampler : Euler a</li><li>CFG Scale : 5-7</li></ul>I highly recommend using the <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">ADetailer</a> extension for better faces and eyes.For v2, I merged v1 with some NoobAI models, @GarejiSama's <a target="_blank" rel="ugc" href="https://civitai.com/models/1649312?modelVersionId=1866796">AdAstra</a> and @Manityro's <a target="_blank" rel="ugc" href="https://civitai.com/models/1732765/vermilion">Vermilion</a>, making v2 much more compatible with NoobAI based LoRAs. The baked in VAE is Anzhc's, which you can get <a target="_blank" rel="ugc" href="https://huggingface.co/Anzhc/Anzhcs-VAEs/blob/main/SDXL%20Anime%20VAE%20Dec-only%20B3.safetensors">here</a>, gives a much sharper look to images, reduced noise, and makes colors less bland overall. LustrousMix v2 generations are more colorful than v1's, with more contrast and sharpness. There is a slight art style change as well, with more mature faces. <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/babdaa02-2197-4a29-9b37-13d2467db4e2/width=525/babdaa02-2197-4a29-9b37-13d2467db4e2.jpeg" /><h3 id="hope-you-enjoy!">Hope you enjoy!</h3>

Files with SHA256: 35c7d7582287f152696874c30364f1bfca122197ae8a01d569cf62dab7489fa9