CivArchive
    Illustrious XL personal merge (noob v-pred0.5 test merge updated) - v3.0_noob-1.0based
    NSFW
    Preview 38131610
    Preview 38131619
    Preview 38131740
    Preview 38132180
    Preview 38141467
    Preview 38141466
    Preview 38141703

    v-pred test merge_normalsteps:

    This version might work fine at 16+ steps, and CFG around 3 (still relatively low CFG).
    these v-pred versions are still test merge, migt not be so stable.
    Normal v3.0 might still be the best version to use.

    このバージョンは16ぐらいのStepsと、3ぐらいのCFG推奨です。
    頭にvp0.5testとついたバージョンはまだテスト段階で、動作に不安定があるかも知れません。
    安定性やクオリティ的には普通のv3.0がまだ一番かと思います。


    v-pred test merge_lowsteps lowcfg:

    This is a personal merge based on noob vpred 0.5.
    While it still is a test merge, I'm uploading it as the results seem not bad.
    First, there are some points to note. Unlike previous EPS versions that were stable with moderate settings, this version is extremely sensitive to Sampler and CFG settings.
    Works nice at low CFG and low steps.

    Recommended Sampler: Euler a SGM Uniform (Not recommending just Euler)
    Recommended CFG: Between 1.0~1.5
    Recommended Steps: 8+ Steps are recommended, but it also operates at a usable level with 4~6 Steps. Try it out for fun.
    Also, this is not necessary but it seems to work nice at 1024x1344 resolution.

    Noob vpred 0.5ベースのテストマージです。
    テストマージではありますが、ある程度満足できる結果にはなりましたので、共有します。
    まず注意点として、既存のPersonal mergeとは違い、SamplerやCFGにかなり敏感です。
    低CFG推奨で、低Stepsでもいい感じの結果を出してくれます。

    推奨Sampler:Euler a SGM Uniform (ただのEulerはおススメしません)
    推奨CFG:1.0~1.5
    推奨Steps:8 Steps以上推奨ですが、4~6 Stepsでも割と使えるぐらいにはなりますので、試してみて下さい。
    また、必須ではありませんが、1024x1344での結果がよさげでした。

    ※ When using the vpred model, the following environments are required:
    Latest dev branch of a1111 webui
    Latest version of forge / reforge
    Latest version of comfyUI

    ※ vpredモデルを使用するにあたって、以下の環境が必要です:
    a1111 webuiの最新のdevブランチ
    forge / reforgeの最新版
    comfyUIの最新版


    v3.0:

    Seems to be quite an improvement for hands and human body, thanks to NoobaiCyberFix fix recipe.(hands and human body from CyberRealistic XL and text encoder CLIP from NoobAI-XL v1.0)
    Of course, thanks to noob 1.0, it also works better for e621-related dirty NSFW stuffs than previous version of personal merge.

    NoobaiCyberFixのレシピを導入したお陰で、手と人体に結構な向上が見られると思います。(CyberRealistic XLから手と人体のデータを取り、TE CLIPはNoobAI 1.0のものを使っています)
    もちろん、noob 1.0のお陰で、e621由来の汚いエロ絵も従来よりうまくこなせます。

    Full recipe:
    Implemented NoobaiCyberFix's fix recipe into 8:4:2:1:1 mix of
    NoobAI-XL v1.0,
    comradeshipXL-v14T13B,
    Obsession v2.1,
    PaSanctuary v2.0,
    The I WonderMix v1


    v2.1:

    Merged 30% of noob-0.75 to v2.0.
    This version is currently my favorite personal merge. While sustaining stable and flat anime style, it's also very capable of e621-related dirty NSFW stuffs such as sc〇t or best〇〇〇ity. (these themes might not be suitable for civitai, so I'm not uploading examples)

    v2.0にnoob-0.75を30%混ぜたものです。
    今んとこ自分用にマージした物の中で一番のお気に入りです。安定的かつフラットなアニメスタイルは維持しながらも、スカ〇ロや獣〇などe621由来の汚いエロ絵もバッチリです。(Civitaiの規約的に怪しいので画像は無しですが)


    v2.0:

    A simple merge of three noob-0.5 based models.
    2:1:1 mix of Comradeship v14T13A, Obsession v2.0 and PaSanctuary V2
    All credit goes to the original creators of each model.


    This is a simple merge model created for personal use by merging three Illustrious XL-based models.

    The Silvermoon, Comradeship, and Copycat models were merged in a ratio of approximately 1:2:2.

    It was originally intended for my personal use, but I'm sharing it because I'm quite pleased with the results.

    All credit goes to the original creators of each model.

    I'm using these quality prompts and negative prompts as below, but it seems to work well even if you remove them or reduce them. Feel free to use them according to your own preferences.

    Quality prompts: masterpiece, best quality, amazing quality, very aesthetic, absurdres
    Negative prompts: lowres, worst quality, bad quality, bad anatomy, jpeg artifacts, signature, watermark

     

    個人的な使用用途で作ったIllustriou XLのマージです。

    Silvermoon、Comradeship、Copycatの3モデルを大体1:2:2ぐらいの比率で混ぜたものです。

    個人使用用途でしたが、なんか結果がかなり気に入ったので共有しておきます。

    元々のモデルの製作者さんたちに感謝いたします。

    自分は以下のようなQuality promptsとNegative promptsを使っていますが、別に適当かいっそ消しちゃっても結果が悪くない気がします。色々とやってみてください。

    Quality prompts: masterpiece, best quality, amazing quality, very aesthetic, absurdres
    Negative prompts: lowres, worst quality, bad quality, bad anatomy, jpeg artifacts, signature, watermark

    Description

    Implemented NoobaiCyberFix recipe(hand and human body from CyberRealistic XL and text encoder CLIP from NoobAI-XL v1.0),
    into 8:4:2:1:1 mix of
    NoobAI-XL v1.0
    comradeshipXL-v14T13B
    Obsession v2.1
    PaSanctuary v2.0
    The I WonderMix v1

    FAQ

    Comments (31)

    DevilLadyNov 3, 2024· 1 reaction
    CivitAI

    I've tried alot of il checkpoints and this one is still the best so far

    LigmaNuts420Nov 4, 2024· 1 reaction
    CivitAI

    Are these Illustrious checkpoints compatible with PONY loras? Would i have to train characters/concepts i want with a particular Illustrious checkpoint?

    dims2
    Author
    Nov 4, 2024· 1 reaction

    Basically not compatible, though some pony loras are partially working at higher weights. It might not work, but you'd better try once.
    Of course, training new ones specifically for Illustrious XL would be much more stable and effective. While I'm not sure which model is particularly good for training, I just use the base Illustrious 0.1 model for lora training.

    LigmaNuts420Nov 4, 2024

    @dims2 thanks for the info. I had no idea they added it as an option on civitai's lora trainer as well.
    SOME loras seem to work just fine, most don't. I think i'm gonna be jumping to Illustrious from Pony now it seems really consistent with both style and anatomical capability.

    472605575Nov 4, 2024
    CivitAI

    When the image is 1536 resolution, the image will be overexposed and the contrast will be particularly high

    Di_Gi_OssanNov 4, 2024
    CivitAI

    Share merges using this model OK とのことで解剖学の改善がとてもよかったのでマージモデルを作らせていただきました。とてもいいモデルですね!😍

    dims2
    Author
    Nov 5, 2024· 1 reaction

    自分も他のモデルをマージにいっぱい使わせて貰ってますので、もちろんOKです!
    というか自分のモデルが他のモデルもマージ材料になるってのも初めての経験?ですので、結構気になりますね笑。自分も使ってみます!

    AlisaaNov 4, 2024· 2 reactions
    CivitAI

    This is hands down one of the best merge of Illustrious on Civitai. Thank you so much for making it.

    meipawsNov 5, 2024
    CivitAI

    Thank you! This is a nice-looking model that has some style to it.

    However, I noticed one issue: the hands seem to have problems in poses like arms at the sides (often generating bad fingers or missing fingers), and even sitting or arm support poses seem to show fewer fingers. The feet are not that great either. Other poses seem to be well-trained; I’m not sure why this is the case.

    Also, can you please generate some sample images without the lora for speed so we know what settings are the best?

    dims2
    Author
    Nov 5, 2024

    Thanks for detailed review! But here's my honest response - that level of fine-tuning is beyond my capabilities. Let's look forward to future improvements of overall illsutrious(and noob) development!

    As for the settings, I should be candid that I'm mostly using them based on intuition, so it's difficult for me to definitively say what's "best." Personally, I see this model not as one that "produces great results with special settings" but rather as one that "consistently delivers decent results with standard settings without much fuss!"

    pyasu620Nov 6, 2024· 1 reaction
    CivitAI

    同アニメの男性主人公キャラカップルが押し流される傾向が強かった

    デフォルト由来の男キャラが強い

    女が男化と竿が増えたり男が増えたりするときも多かった

    結構弊害もあるようだ

    その代わりこのデータに支えられるポーズが補強される感じ

    私もテストでマージしてみたけどあまりうまくいかなかった

    *追加調査したらv1の傾向でもあるようだった

    vpredを今は深く見ています

    low_channel_1503Nov 7, 2024
    CivitAI

    probably my favorite merge right now, looking forward to see if you can do a vpred version?

    dims2
    Author
    Nov 11, 2024

    I'm testing some merges after noob vpred 0.5, but noob vpred 0.5 itself is quite fine, and my vpred test merge doesn't feel satisfiying enough so I'm not sure I'll upload it.

    low_channel_1503Nov 11, 2024

    @dims2 i see. i was hoping to see it to get some better anatomy, since it breaks much more often than your merge for certain prompts, but i understand if you wont upload it

    seawolf338Nov 7, 2024· 12 reactions
    CivitAI

    the strongest model to date.

    1) definitely stronger than PDXL

    2) perfect adherence to the prompt

    3) flexible graphics settings through artists

    4) perfect anatomy even without a negative prompt.

    dear author, I hope you will not stop at the achieved success and will please us with new models!

    r_ainNov 7, 2024
    CivitAI

    Really great model!

    LigmaNuts420Nov 7, 2024· 1 reaction
    CivitAI

    Will this get more updates? I've tried i think 6 different illustrious models so far, and your model seems to be by far the best in every metric.

    Stocking Anarchy is the main character i've used to benchmark models since Pony Diffusion and yours is the only one that consistently got every detail almost perfectly with every gen, only problem is eye accuracy, seems to get kinda noisy/grainy/broken every other gen and depending on the style. Really hoping this gets further ironed out. Gonna post some of those gens in a bit today. Wanted to throw this feedback out there because i'm floored how good this model is.

    dims2
    Author
    Nov 11, 2024· 1 reaction

    Hi, thanks for review! It will get more updates if I feel satisfied for new merges.
    I'm currently testing some merges after noob vpred 0.5, but noob vpred 0.5 itself is quite fine enough, and my vpred test merge doesn't feel satisfiying enough so I'm not sure I'll upload it now.

    LigmaNuts420Nov 11, 2024

    @dims2 cool, thanks for the info. Seems illustrious/noob is going to be updated pretty frequently throughout the remainder of the year so its nice to know you're on top of it.

    waterdrinkerNov 9, 2024
    CivitAI

    ⭐⭐⭐⭐⭐

    liftweightsNov 11, 2024· 2 reactions
    CivitAI

    Insane model. Works really well with DMD2 distillation at 8 steps.

    dims2
    Author
    Nov 11, 2024· 1 reaction

    I was still using Hyper 8Step, but after seeing this comment, I tried DMD2 for the first time, and it seems to have much better compatibility. I'll likely switch to DMD2. Thanks for info!

    low_channel_1503Nov 11, 2024

    do you have an example of a workflow with dmd2?

    z0t070201Nov 13, 2024· 1 reaction
    CivitAI

    It is a very nice checkpoint with few human body breaks.

    I would like to borrow your knowledge, but the square where the scene description is written is appearing unintentionally, and I am struggling to find a solution to this problem.

    What can I enter in the negative prompt to suppress this appearance?

    kirakiraNov 13, 2024
    CivitAI

    エロいお兄さんにはこのモデルホントおすすめ

    aimaliasNov 14, 2024· 2 reactions
    CivitAI

    Works great when you include the character name + series name + artist name in the prompt, optionally with official art and/or anime screenshot.

    DivinesNov 15, 2024
    CivitAI

    Can someone tell me which loras will work, do i need to search for Noob and illus loras now?

    seawolf338Nov 15, 2024· 1 reaction

    only illus loras

    DivinesNov 15, 2024

    @seawolf338 Thanks

    dims2
    Author
    Nov 15, 2024· 1 reaction

    @Divines For v1.0, illus loras will not work very well, but for v2.0+, both illus and noob loras will work fine.

    DivinesNov 15, 2024

    @dims2 Using v3 so i can use both alright noted ^^

    Checkpoint
    Illustrious

    Details

    Downloads
    33,873
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/3/2024
    Updated
    5/13/2026
    Deleted
    -

    Files