かなり低品質です、期待しないで下さい。
It's quite low quality, so don't expect too much.
*********************************
衣装を変更する為にタグ固定していません。
表示させたい部分以外は消して使用して下さい。
例、
cowboy shotなら、boots,を消す
目を閉じたいならblack eyes,を消してclosed eyes, を追加
など。
The tag is not fixed to change the costume.
Please erase the part other than the part you want to display.
example,
For cowboy shot, remove boots,
If you want to close your eyes, remove black eyes, and add closed eyes,
Such.
*********************************************
weight 0.7、ADetailerを使用しface_yolov8nで顔修正を行っています。
以下、トリガー、キャラ容姿、上衣装、下衣装
Weight 0.7 ADetailer is used and face_yolov8n is used for face correction.
Below, trigger, character appearance, upper costume, lower costume
*********************************************
beanbandit,1man,
short hair,black hair,bangs,scar,brown eyes,chin,muscle,tall man,
bandana,
leather jacket,shirt,open jacket,sleeves rolled up,
fingerless gloves,
jeans,
shoes,
Description
キャラテスト
まったくよらない、傷も覚えない。
FAQ
Comments (9)
何度となくやり直したが、まったく上手くいかない、たまにこんな題材あるよね・・
とりあえずBUFF入れてごまかしてみたｗ
I see you are very fast in doing Lora!
Would you please some Lora about Urusei Yatsura?
Actualy, on CivitAI there are some "Lum", one "Benten" and two "Ran"... but is missing: "Kurama", "Mendo" and "Ataru".
Unfortunately, I can't make it because I don't have the "material picture".
残念ながら"素材絵"を持っていないので作れません。
@jibunsagasinotabi I can watch the episodes, grab images and pass images to you.
@jibunsagasinotabi I have already did it for first 5 episodes, grabbing the best images for Ataru and Mendo. (It took several days. )
Let me know if this can be a good dataset for you, or you need more or different images.
For example, having a full body image is very rare in this anime... is this a problem?
Ataru raw images:
https://mega.nz/file/6hcHBSRD#QJaPnB_DFhi2PtHl4NL1HsZI-CPGEzHGwEcxuNQcTK4
Mendo raw images:
https://mega.nz/file/ullHTZIQ#jRrs_QcJZj5jNaMUgI1-9nLiZmUNFSEgQiq4bORxCdo
@settima_ai I checked the "Material Pictures". There are only a few materials that can be used. Without a full-body picture, it is often impossible to express the character's whole body. The next problem is that you can only use one of the images with the same background (different facial expressions in one scene), so you need to select an image. What I can create from this material picture will be a drawing of the upper body of the character, is that okay?
There is only one photo of the same pose that can be used.If costumes are required, a full-body image of each is required, but if the costumes are in the model data, they are not necessary.
"素材絵"確認しました、使える素材は少数です、全身絵が無いと、キャラクターの全身を表現出来ない事が多いです。次の問題は、背景が同じ物（1シーンで表情が違うもの）はその中の1枚しか使えません、画像を選択する必要があります。この素材絵から出来るものは、キャラクターの上半身を描く物になるでしょう、それでも良いですか？
使える、同じ姿勢も1枚です、衣装が必要な場合は、各、全身像が必要ですが、モデルデータ内に有る衣装なら必要ありません。
@jibunsagasinotabi No, don't do a LORA with a "partial" chararter. Please wait until I search for better images and full body pictures.
I made "same scene with different facial expressions" because I was believing it will be usefull to teach SD how that character is smiling, sad, angry etc... but now I understand this is not usefull.
So, i'm reducing the "duplicated" images while i'm searching new ones. It will take some days, please stand by.
@settima_ai とりあえず、今ある素材だけで作成したので見て下さい。私は何か違う気がしてます。
For now, I created it using only the materials I already have, so please take a look. I feel like something is different.
@jibunsagasinotabi Wow! You are faster then light! Many thanks, I will soon test them!
Yes, expecially Ataru resembles correctly in just only one image of yours. In the other two, It's a bit "off", like you said before. But let me test it...
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.