Starting from V3, it's very important to use (realistic, 3d:1.2 - 1.4) in your positive prompt for the model to work as intended. I prefer to set weight to 1.3.I've tried to merge an anime model that will look as detailed as possible, yet remain vivid and clean, with no realistic features (nose, nostrils) on character's face. Also I've tried to make character's skin realistic'ish: soft and smooth. Prompt versatility is also a must for me. So for now I came up with this merge. That's my first time using SuperMerger, I still don't know a lot of things about what block does what, but I did my best at the moment. I hope you'll like my mix. That's not a final version though, I'll work on updating it in the future.All preview Images made with Hires and ADetailer.Recommendations:<ul><li>Use danbooru tagging for prompts, <a target="_blank" rel="ugc" href="https://github.com/DominikDoom/a1111-sd-webui-tagcomplete">THIS</a> extension will help you with that;</li></ul><ul><li>I'm using <a target="_blank" rel="ugc" href="https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/main/vae/kl-f8-anime2.ckpt">kl-f8-anime2 VAE</a>, but if you find it too colorful, feel free to use any of your favourite ones;</li><li>Set Clip Skip to 2, it's very important for anime models to work properly;</li><li>Samplers: Euler A, 30 steps, for fast, stable and cleanest generations; DPM++ SDE, 24 steps for more details, but generation time will increase significantly;</li><li>Upscaler: <a target="_blank" rel="ugc" href="https://huggingface.co/Akumetsu971/SD_Anime_Futuristic_Armor/resolve/main/4x_fatal_Anime_500000_G.pth">4x_fatal_Anime_500000_G</a>, is the best IMO, Upscale by 2, 15 steps for Euler A, 10 steps for any other sampler, 0.3-0.4 denoising strength;</li><li>CFG Scale: 6-8, my preferred scale is 6.5;</li><li>Try <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">ADetailer</a> extension for drastically better faces. I never generate without it, it's just too good;</li><li>For prompt base I'm using <code>(best quality, absurdres:1.2)</code> in positive, and <code>(worst quality, low quality:1.2), (blurry, bokeh, dof, fog:1.2)</code> in negative;</li><li>Use <a target="_blank" rel="ugc" href="https://civarchive.com/models/58726/asyncs-negative">Async's Negative</a>, or <a target="_blank" rel="ugc" href="https://civarchive.com/models/7808/easynegative">EasyNegative</a> embedding, I prefer to put it on 0.8 weight. The first one will give you more sharp, contrast image with slightly better hands. The second one is softer, with more volumetric lighting. Remember not using it with Style LoRAs, because it may ruin the style;</li><li>You can use <a rel="ugc" href="https://civarchive.com/models/16993?modelVersionId=20068">banhandv4</a> for better hands, it's the best embedding for hands I found;</li><li><s>Also try</s> USE<code>realistic, 3d</code> tags, those are important for the model to work as intended. I recommend setting weight to from 1.2 to 1.4. I prefer 1.3 for my generations;</li></ul>Credits:<ul><li>Huge thanks to <a target="_blank" rel="ugc" href="https://civarchive.com/user/TroubleDarkness">TroubleDarkness</a> for teaching me everything I know about Stable Diffusion. Go check out his <a target="_blank" rel="ugc" href="https://civarchive.com/models/6174?modelVersionId=432048">DivineElegance</a>, it's amazing;</li><li>Thanks to <a target="_blank" rel="ugc" href="https://civarchive.com/user/Kybalico">Kybalico</a> for making <a target="_blank" rel="ugc" href="https://civarchive.com/models/105133/strawberryshortcake">StrawberryShortcake</a>;</li><li>Thanks to <a target="_blank" rel="ugc" href="https://civarchive.com/user/technobyte">technobyte</a> for making <a target="_blank" rel="ugc" href="https://civarchive.com/models/130417/milky-wonderland">Milky Wonderland</a>;</li><li>Thanks to <a target="_blank" rel="ugc" href="https://civarchive.com/user/newlifezfztty761">newlifezfztty761</a> for making <a rel="ugc" href="https://civarchive.com/models/28169?modelVersionId=138620">CuteYukiMix</a>;</li><li>Thanks to You, for trying my mix. Have fun!</li></ul>

RyoMix

(I updated not only model, but some recommendations too, go check them out.) I reverted back to V1 for making a new merge. In this version I and added glorious <a rel="ugc" href="https://civitai.com/models/28169?modelVersionId=138620">CuteYukiMix</a>, one of my favourite models. This should help with hands, anatomy and stability.

ryomix_v30.safetensors

v3.0

v2.0

v1.0

creator: RyoidenshiI've tried to merge an anime model that will look as detailed as possible, yet remain vivid and clean, with no realistic features (nose, nostrils) on character's face. Also I've tried to make character's skin realistic'ish: soft and smooth. Prompt versatility is also a must for me. So for now I came up with this merge. That's my first time using SuperMerger, I still don't know a lot of things about what block does what, but I did my best at the moment. I hope you'll like my mix. That's not a final version though, I'll work on updating it in the future.All preview Images made with Hires and ADetailer.Recommendations:<ul><li>Use danbooru tagging for prompts,&nbsp;<a href="https://github.com/DominikDoom/a1111-sd-webui-tagcomplete" rel="ugc noopener noreferrer nofollow" target="_blank">THIS</a>&nbsp;extension will help you with that;</li></ul><ul><li>I'm using&nbsp;<a href="https://huggingface.co/hakurei/waifu-diffusion-v1-4/blob/main/vae/kl-f8-anime2.ckpt" rel="ugc noopener noreferrer nofollow" target="_blank">kl-f8-anime2 VAE</a>, but if you find it too colorful, feel free to use any of your favourite ones;</li><li>Set Clip Skip to 2, it's very important for anime models to work properly;</li><li>Samplers: Euler A, 30 steps, for fast, stable and cleanest generations; DPM++ SDE, 24 steps for more details, but generation time will increase significantly;</li><li>Upscaler:&nbsp;<a href="https://huggingface.co/Akumetsu971/SD_Anime_Futuristic_Armor/resolve/main/4x_fatal_Anime_500000_G.pth" rel="ugc noopener noreferrer nofollow" target="_blank">4x_fatal_Anime_500000_G</a>, is the best IMO, Upscale by 2, 15 steps for Euler A, 10 steps for any other sampler, 0.3-0.4 denoising strength;</li><li>CFG Scale: 6-8, my preferred scale is 6.5;</li><li>Try&nbsp;<a href="https://github.com/Bing-su/adetailer" rel="ugc noopener noreferrer nofollow" target="_blank">ADetailer</a>&nbsp;extension for drastically better faces;</li><li>For prompt base I'm using&nbsp;(best quality, absurdres:1.2)&nbsp;in positive, and&nbsp;(worst quality, low quality:1.2), (blurry, bokeh, dof, fog:1.2)&nbsp;in negative;</li><li>Use&nbsp;<a href="https://civitai.com/models/7808/easynegative" rel="ugc noopener noreferrer nofollow" target="_blank">EasyNegative</a>&nbsp;embedding, I prefer to put it on 0.8 weight. Remember not using it with Style LoRAs, because it may ruin the style;</li><li>You can use&nbsp;<a href="https://huggingface.co/Cordeliya/animefull-latest/resolve/main/bad-hands-3.pt" rel="ugc noopener noreferrer nofollow" target="_blank">Bad Hands 3</a>&nbsp;for better hands, just don't rely on it too much;</li><li>Also try&nbsp;realistic, 3d&nbsp;tags, those are really good. Be careful with weight though;</li></ul>Credits:<ul><li>Huge thanks to&nbsp;<a href="https://civitai.com/user/TroubleDarkness" rel="ugc noopener noreferrer nofollow" target="_blank">TroubleDarkness</a>&nbsp;for teaching me everything I know about Stable Diffusion. Go check out his&nbsp;<a href="https://civitai.com/models/6174?modelVersionId=432048" rel="ugc noopener noreferrer nofollow" target="_blank">DivineElegance</a>, it's amazing;</li><li>Thanks to&nbsp;<a href="https://civitai.com/user/Kybalico" rel="ugc noopener noreferrer nofollow" target="_blank">Kybalico</a>&nbsp;for making&nbsp;<a href="https://civitai.com/models/105133/strawberryshortcake" rel="ugc noopener noreferrer nofollow" target="_blank">StrawberryShortcake</a>;</li><li>Thanks to&nbsp;<a href="https://civitai.com/user/technobyte" rel="ugc noopener noreferrer nofollow" target="_blank">technobyte</a>&nbsp;for making&nbsp;<a href="https://civitai.com/models/130417/milky-wonderland" rel="ugc noopener noreferrer nofollow" target="_blank">Milky Wonderland</a>;</li><li>Thanks to You, for trying my mix. Have fun!</li></ul>

(I updated not only model, but some recommendations too, go check them out.) I reverted back to V1 for making a new merge. In this version I and added glorious CuteYukiMix, one of my favourite models. This should help with hands, anatomy and stability.


Files with SHA256: 7f707597b85b5f2a865b3ad77707241ca54cba3a783f76705debb886d88e2c3c