Chilloutmix

*I modified the License.(22/2/Feb) That is not usual "creativeml-openrail-m" Check Permission and Liscense below(modified Dreamlike Liscense.) When I made samples, My eta (see setting) was "0". And ESND was "1" Now,I canged them to eta to "0.67" and ESND "31337"(20/Feb/2023) よいこのみんなは節度を守って楽しもうな！！ 職場や家族の前で開くのはやめておいたほうがいいよ！！ 当モデルを転売することは認めていません。 生成された画像の取り扱いには、自己責任の上、十分注意してください。 当モデルを使用したことを明記せずに生成された画像を公開することは禁止です。 商用利用は一切認めていません。 AIコミュニティの発展に悪影響を与えるような行動や個人の権利やプライバシーを侵害するような利用は禁止します。 New merge model for 2.5D illustration here!! ↑ (少しイラスト寄りにしたのもあります。) IMPORTANT:First of all, I never suggest regenerate images of “real “person, but, photo "realistic" images. 本人郑重声明：本模型原则上禁止用于训练基于明星、公众人物肖像的风格模型训练，因为这会带来争议，对AI社区的发展造成不良的负面影响。如各位一定要违反以上声明训练相关模型并公开发布，请在您的发布说明中删除与本模型有关的一切描述。感谢各位使用者的支持与理解。 I solemnly declare: In principle, this model is prohibited from being used for training style models based on portraits of celebrities and public figures, because it will cause controversy and have a negative impact on the development of the AI community. If you must violate the above statement to train the relevant model and release it publicly, please delete all descriptions related to this model in your release notes. Thank you for your support and understanding. I appreciate you enjoy my model. But, it might cause legal conflicts if you made some works/Loras/Embeddings named with actual person/copyright. I beg you never to put them with my model. If you can't be with actual name/copyright, just remove my model from your works. I never suggest to regenerate actual persons/copyright with my models. And I never want any legal conflicts from my models. PLEASE!! Care about legal conscious and privacy!! Show what your favorite works from this merge on comments, if you like this mix. I wanna see and share your good ideas from this merge. :) It’s my pleasure. Let's share your good ideas and enjoy this merge!! (Some funs have already put their great works from this merge. It could be like a kind of community.) I started to put some tips with samples/prompts on review. Try to find them!!(I've already put 5. 14/Feb/2023) (I'm sorry but I stop to invite my discord.) Konnichiwa!!!! ・This is Merged "Basilmix"( + wonderful realistic models. ( by ) ・Use: vae-ft-mse-840000（←！！） ) 1. Get vae-ft-mse-840000-ema-pruned.safetensor 2. Put it in folder(models>VAE) 3. Set VAE on webUI(Settings>Stable diffusion>SD VAE) 4. Don't forget to push "Apply settings" (Your picture would be foggy, if you didnt use vae with this merge.) You dont need to use VAE with the latest ver."Chilloutmix-Ni" ・All sample images have been upscaled using highrexfix And, I made some samples when I made test ckpt file for myself. I made safetensor file after that. (Cuz, ckpt files couldnt be liked on this site, because they include arbitrary code.) That’s why Prompts from some sample pictures wouldn't make the same pictures,sorry. ・I think this model is not perfect, but better for realistic Asian(+ other areas.Just try) . ・Face variation is really limited. Use LORA or Embeddings as you like.(←！！) I like to use "Ulzzang-6500" embeddings (I used that for all sample images). This is my type....I couldn't avoid generating with this.... You can get it here. (Thanks and for telling me on comments.) ↓ (all 3 adress are the same files) Other Embeddings By By Dcy:Asian face mix (Thanks for sharing !!) As the same as other areas of this world, Asian are quite variated and there are quite lots of races, communities and nations in Asia. Who can tell you what is the typical Asian? '"Typical "could cause less variation. I've been coming to think face variation should be developed not on models, but on LORAs and also on embeddings. Because,results can be quite randomed if that was developed on models. (Yeah,I know random encounting could be also fun,though.) You can make ideal faces with LORAs, embeddings as you like. (Some examples of face variation by using some TI/HN/LORA) Some Lora could work very nice, even though they were made for 2D charas. They are notable that 2D charas Loras generate not only crothes, hair styles, but also facial features!! This model is likely to get bigger tits. If you like more smaller tits, try both of prompts, "(small breasts) and (fantasy breasts)" (Slim girl) maybe good. ・Preparing: I put explanation too much,sorry. ーーーーーーーーーーーーーーーーーーーーーーーーーーーーーーーー Due to using , this model has the following license: Modified License（Dreamlike photoreal allowed me to modify Liscence.） This model is licesed under a modified CreativeML OpenRAIL-M license. - You can't host or use the model or its derivatives on websites/apps/etc., from which you earn, will earn, or plan to earn revenue or donations. If you want to, please email us at - You are free to host the model card and files (Without any actual inference or finetuning) on both commercial and non-commercial websites/apps/etc. Please state the full model name (Dreamlike Diffusion 1.0) and include a link to the model card ( - You are free to host the model or its derivatives on completely non-commercial websites/apps/etc (Meaning you are not getting ANY revenue or donations). Please state the full model name (Dreamlike Diffusion 1.0) and include a link to the model card ( - You are free to use the outputs of the model or the outputs of the model's derivatives for commercial purposes in teams of 10 or less （Edit:We sencerely ban to use the outputs of the model or the outputs of the model's derivatives(further murged models and Loras included) for commercial purposes. Because, it would cause controversy and have a negative impact on the development of the AI community, wheter images were generated from real person or not.） - You can't use the model to deliberately produce nor share illegal or harmful outputs or content - The authors claims no rights on the outputs you generate, you are free to use them and are accountable