This LoRa has learned JAV of Hard Deep Throat Play.
❤This is a concept model for hardcore deep throat play, with her pigtails held in her hands.❤
This LoRa training material was created using the following method
Real‗LoRa was created using frame capture material from a JAV video.
The training material from Real_LoRa was replaced with Anime material using img2img batch processing, again using the originally created Real_LoRa.
Square 1024x1024 ......... ❤️Highly recommended❤️
Landscape 1216x832 .... ❤️Highly recommended❤️
Portait 832x1216 ............ 👍It feels great! 👍
The learning materials are available here. Please feel free to use them.
https://mega.nz/folder/BhoilbDL#A61KbnabypK2Utb-HpjlEA
--SAMPLE PROMPT ------------------------------------------------------------------------------------
PROMPT:Base
<lora:JAV_MIZD-490_DeepThroatHandle_Anime_IL_V2:1.3> mizd-490, 1girl, twin tails, 1boy, penis, hetero, male pubic hair, perfect hands,
(grabbing another's hair (twin tails) (with both hands) ), oral, fellatio, irrumatio, deepthroat, saliva, solo focus, crying, blush, string sweat, string snot,
<***Character Prompt***>.
--Sorry Japanese Text ----------------------------------------------------------------------------------
このLoRaは、イラマハンドルのオーラルセックスをコンセプトにしています。
ＪＡＶの説得力のある画面構成での、「ツインテール両手握りのイラマチオハンデル」 をコンセプトにハードフェラチオ画像をひたすら垂れ流してください。
1次学習元のアダルトビデオの実写映像は、全てHENTAI ANIMEに置換・最適化されて２次学習されています。オリジナルのアダルトビデオに由来する実写画像はこのLoRaでは生成されません。
学習素材はこちらです。ご自由にお使いください。
https://mega.nz/folder/BhoilbDL#A61KbnabypK2Utb-HpjlEA
このLoRaトレーニング教材は、以下の方法で作成されました。
JAV動画のフレームキャプチャ素材を作成し、まずReal_LoRaが作成されました。
img2img のBatch処理で①で作成したReal_Loraも活用し、Real_LoRaのトレーニング教材を不特定キャラクターのAnime素材に置換しました。
優良素材を抽出し、JAV Video由来のHentai Anime LoRaの学習素材が完成しました。
縦長では単調な生成になりがちです。正方形か横長でのダイナミックな生成をオススメします。
サンプルスクリプトの <***Character Prompt***> 部分にお好きなキャラクターを指定してご使用ください。
私は本LORAの利用に、制限・権利・責任を一切主張しません。
Description
FAQ
Comments (8)
綺麗にセットされた女性の美しい髪の毛を引っ張っての、乱暴なオーラルプレイをする動画が大好きです。
涙と涎で、派手にメイクが崩れて顔面がドロドロになっちゃうのは、リアル動画だとすごくイイコンセプトなのですが、アニメヘンタイだとなんか違うなー？ってなりましたので、今回のLoRaでは見送りました。
Works pretty good with realism checkpoints and characters too (though the grab doesn't look as strong and intense as with anime pics), love it! Great work as always
Thank you for your comment!
The first stage of creating this LoRa was created from a realistic image, so naturally it is for real-world use.
Previously, I had released materials and LoRa for real-world use, but it seems that they would be subject to CivitAI censorship, so I am currently refraining from releasing the materials for real-world use.
Curious about the training.
How do you do Image to image from real to anime with Illusterious? (Have not actually learned how to train Illusterious either, is there a good guide you can recommend?)
Thanks for your comment.
My LoRa development environment is entirely local.
I don't use CivitAI's online learning environment.
I use the batch processing feature of StableDiffusion's "img2img" function to batch process live-action frame capture images and convert them into animated images.
I may write an article on guidelines for generating LoRa images from video in the future.
@suteakasuteakasuteka434 FI you do write a guide i would be keen to read it (I havent trained lora's in a while, but all were done local, I have a 3090. Using flux gym....which....well...only effective for flux, which has somewhat fallen out of favour these days.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.