PixAI Üyeliği
PixAI
PixAI
Generate
Sign in
Ana Sayfa
Oluşturma
Modeller
Studio
NEW
Yarışma
Mio.2 ile oluştur
Araç Kutusu
NEW
Guide

SDXL

Nonnette / iL

#Plastic model
Yapay Zekâ Araçları
  • Yapay Zekâ Anime Üretici
  • Yapay Zekâ Görüntü Oluşturucu
  • Yapay Zekâ Animasyon Oluşturucu
  • Anime Figür Üretici
  • Karakter Kartı Üretici
  • Dergi Kapağı Üretici
  • Peluş Oyuncak Üretici
  • Dev Heykel Üretici

Keşfet

  • Trend Etiketler
  • Sıralama
  • Model Pazarı
  • Yarışma
  • Haberler

Hakkında

  • Guide
  • PixAI Nasıl Kullanılır
  • Tsubaki.2
  • Mio ile tanışın
  • İçerik Kuralları

Fiyatlandırma ve Yardım

  • Üyelik
  • Kredi Paketleri
  • İletişim

Mobil Uygulama

  • PixAI uygulamasını edinin
  • App Store
  • Google Play
© 2026 PixAI
  • Hizmet Şartları
  • Gizlilik Politikası
  • Telif Hakkı Politikası
  • Attributions
  • Belirtilen Ticari İşlemler Kanunu
Nonnette / iL - AI Model cover image
Yüklenme Tarihi
21 Şub 2025 13:49
Kullanım
12.2k
Yorumlar
Mükemmel  (1)
Örnekleme Adımları

25

Örnekleme Yöntemi

Euler a

CFG Ölçeği

7

Negatif

lowres,bad quality, worst quality,bad anatomy,sketch,jpeg artifacts,ugly,poorly drawn,blurry,watermark,simple background,transparent background,

İzinler
  • Allow image generation & sharing
  • Kullanıcıların modelinizi indirmesine izin verin
  • Ticari kullanımlar

Recommended LoRAs

Shiny_Niji | Style_Yoon.A - AI Model cover image

User LoRA

Shiny_Niji | Style_Yoon.A
72
(改良調整中)強化ツール_高精細・カラーリング・ディテール - AI Model cover image

User LoRA

(改良調整中)強化ツール_高精細・カラーリング・ディテール
33
ahriii artist artstyle - AI Model cover image

User LoRA

ahriii artist artstyle
787
Urakagi Saji Style for Illustrious - AI Model cover image

User LoRA

Urakagi Saji Style for Illustrious
36

Açıklama

日本語解説は下方にありますNonet is a miscellaneous Noob-based Eps merge model that can be found anywhere.Basically, if you follow the usage of Illustrious eps, you can use it without any problem.・v010 is a base model, so it is really a miscellaneous merge model that can be found anywhere.・v020 is a fine tuning of v010 with lametta images, and then multiplied by a merge model with Lora applied, which is also created from lametta images, to get the output I like.v020 does not need the negative prompt “bright pupils, white pupils,” which I recommended before. It is probably cleaner not to include it unless there is a problem.・v034 inherits the merging method of v020, but has been refined by reviewing the material.・v036 I felt that v034 was a little out of adjustment, so I remerged it and added a little " featureless25DMix_v10 ".How toSampling method Euler a or DPM++ 2M (You may be able to use something else (I haven't tried it))Sampling steps >20CFG Scale >3 (The adjustment was made in 5)recommended that the Hires. fix be implemented.ESRGAN:Denoising strength About 0.4~0.5Latent:Denoising strength 0.6 and aboveIt is still easy to increase the number of fingers due to lack of adjustment. I don't know what to do. hummm…As a recommended promptQuality prompts samplemasterpiece,amazing quality,very aesthetic,absurdres,newest, (Adjust to your taste)Negative prompts samplelowres,bad quality,worst quality,bad anatomy,sketch,jpeg artifacts,ugly,poorly drawn,blurry,watermark,simple background,transparent background, (Adjust to your taste)merged recipe・v010lilybreed_v10 + (kiwimixXL_v3 - daisybreed_v10) = AA + catCarrier_v20 = BcatTowerNoobaiXL_v10 + whynotIlluWainoob_v30 = CillustriousXLPersonalMerge_v30Noob10based + sakuraplatinum_v20241117 = DobsessionIllustrious_v30 + holyMixIllustriousxl_v1 = Edaisybreed_v10 + dreammix_v10 = FC + D + E + F = GB + G = Nonnette_v010VAE: liquid111vae_sdxl9745・v020Fine tuning nonnette_v010 with 520 lametta output images = AntrMIXIllustriousXL_xii + rillusmRealistic_v11 = BMerging B with Lora created from 60 lametta output images = Csakuraplatinum_v20241117 + waiNSFWIllustrious_v70 = DMerge Lora created from 60 lametta output images into D = EA + C = FA + E = GF + G = nonette_v020VAE: liquid111vae_sdxl9745・v034Fine tuning Rillusm Realistic Illustrious v1.1 with 600 lametta output images = AFine tuning exploits_v1 with 600 lametta output images = BFine tuning catTowerMIX with 600 lametta output images = CA + B + C = DD + mergeijPONYIL_ilV10VAE = EE + B = Nonnette_v034VAE: xlVAEC_f2・v036Fine tuning catTowerMIX with 600 lametta output images = A (Same as C in v034)nonnette_v034 + A = BB + featureless25DMix_v10 = nonnette_v036VAE: xlVAEC_f2Thanks to the authors of each of these wonderful models !!ノネットはどこにでもあるような雑多なNoob系Epsマージモデルです。基本的にはIllustrious eps の使い方に従ってもらえれば問題なく使用できると思います。このモデル標準の絵柄にIllustrious(Noob)が内包しているデータをうまく組み合わせて使用することを前提としてマージしてあります。特定のキャラクターなどを呼び出してといった使い方は他のモデルに及びません。・v010はベースモデル的なやつなので本当にどこにでもあるような雑多なマージモデルです。・v020はv010をlamettaの画像でFine tuningしたものに、同じくlamettaの画像から作成したLoraを当てたマージモデルをかけ合わせて私好みの出力となるようにしました。v020は以前推奨したネガティブプロンプトの「 bright pupils, white pupils, 」は基本不要です。問題がない限り入れない方が多分きれいです。・v034はv020のマージ方法を継承しつつ素材を見直ししてリファインしました。・v036 v034が少し調整不足に感じたので再マージを行うとともに「featureless25DMix_v10」をちょい足ししてあります。使い方サンプリングはEuler a か DPM++ 2Mが推奨で、モデル調整は主にEuler aで行いました。他のサンプラーは試していませんが何かしらの絵は出ると思います?CFGは3~7あたりが推奨で、調整は5と7で行っていました。Hires. FIXするとより良い出力が得られます。この時ESRGAN系(4x-AnimeSharpとかR-ESRGAN 4x+ Anime6Bなど)はDenoising strengthを0.4~0.5付近にするのがおすすめです。Latent系なら0.6~0.7あたりが良いと思います。調整不足からかやっぱり指は増えやすいですね。どうしたものか。推奨プロンプトクオリティプロンプトとしてmasterpiece,amazing quality,very aesthetic,absurdres,newest,あたりは入れておいてください。適時調整してね。ネガティブプロンプトとしてlowres,bad quality,worst quality,bad anatomy,sketch,jpeg artifacts,ugly,poorly drawn,blurry,watermark,simple background,transparent background,こんな感じで、もちろん適時調整してね。マージレシピ・v010lilybreed_v10 + (kiwimixXL_v3 - daisybreed_v10) = AA + catCarrier_v20 = BcatTowerNoobaiXL_v10 + whynotIlluWainoob_v30 = CillustriousXLPersonalMerge_v30Noob10based + sakuraplatinum_v20241117 = DobsessionIllustrious_v30 + holyMixIllustriousxl_v1 = Edaisybreed_v10 + dreammix_v10 = FC + D + E + F = GB + G = nonnette_v010VAE: liquid111vae_sdxl9745・v020nonnette_v010をlamettaの出力画像520枚でFine tuning = AntrMIXIllustriousXL_xii + rillusmRealistic_v11 = BBにlamettaの出力画像60枚から作成したLoraをマージ = Csakuraplatinum_v20241117 + waiNSFWIllustrious_v70 = DDにlamettaの出力画像60枚から作成したLoraをマージ = EA + C = FA + E = GF + G = nonette_v020VAE: liquid111vae_sdxl9745・v034Rillusm Realistic Illustrious v1.1をlamettaの出力画像600枚でFine tuning = Aexploits_v1をlamettaの出力画像600枚でFine tuning = BcatTowerMIXをlamettaの出力画像600枚でFine tuning = CA + B + C = DD + mergeijPONYIL_ilV10VAE = EE + B = Nonnette_v034VAE: xlVAEC_f2・v036catTowerMIX を600枚のlamettaに出力によりFine tuning = A (v034のCと同じもの)nonnette_v034 + A = BB + featureless25DMix_v10 = nonnette_v036VAE: xlVAEC_f2上記の素晴らしいモデルを作成してくださった方々に感謝いたします。どうもありがとう!

Yorumlar