アニメLORAですが、アニメモデルとの相性は良くありません、"kantanmix"以外のモデルで出力する場合は2.5Dモデルで試してみて下さい。アニメモデルの場合は、重量を下げ、プロンプトで補正したりする事が必要です。
Although it is an anime LORA, it is not compatible with anime models. If you want to output with a model other than "kantanmix", please try using a 2.5D model. For animated models, it may be necessary to lower the weight and make corrections with prompts.
*********************************************
好きな衣装や、背景、ポーズを組み合わせて下さい。
Please combine your favorite costumes, backgrounds, and poses.
*********************************************
weight 0.7、ADetailerを使用しface_yolov8nで顔修正を行っています。
高解像度補助:R-ESRGAN 4x+ Anime6B,倍率2倍
Weight 0.7 ADetailer is used and face_yolov8n is used for face correction.
Hires.fix:R-ESRGAN 4x+ Anime6B,2x magnification
*********************************************
"Detail Slider - LoRA"
https://civarchive.com/models/153562?modelVersionId=171989
"Detail Tweaker LoRA (细节调整LoRA)"
https://civarchive.com/models/58390?modelVersionId=62833
"flatBG"
https://civarchive.com/models/81298/flatbg
"noline"
https://civarchive.com/models/81309/noline
"boldline"
https://civarchive.com/models/81313/boldline
"Zoom Slider - LoRA"
https://civarchive.com/models/114460?modelVersionId=123732
"ADetailer"
(インストールは自己責任でお願いします)(Please install at your own risk.)
顔の詳細を修正する追加機能
Additional features to correct facial details
https://github.com/Bing-su/adetailer
Embeddings
"Fast Negative Embedding (+ FastNegativeV2)"
https://civarchive.com/models/71961?modelVersionId=94057
"EasyNegative"
https://civarchive.com/models/7808/easynegative
"negative_hand Negative Embedding"
https://civarchive.com/models/56519/negativehand-negative-embedding
Description
キャラ固定テスト
FAQ
Comments (8)
体の入れ墨学習させるの忘れてたｗ
まぁ、スクナを脱がせて喜ぶ人はいないでしょｗ
I have a character I want you to make!
Will you make it?
I can't answer that question unless I know who it is.
それが誰なのかが判らないと答えようがない。
@jibunsagasinotabi Sorry, I guess I asked the wrong question.
Can you tell me how you LORA-ize like this?
How do you model the various characters?
@tickling
kohyaSSを使用してDreamboothLOLAを使用しているのが前提です。
作りたいキャラクターの高品質な教師画像があれば比較的簡単に出来ます。
Train batch size,Epoch,は各教師画像を750～1000回学習する様調整
Mixed precision、Save precision、は使用しているGPUに合わせて設定
LR Scheduler,Optimizer,Adafactor
Network Rank 64,Network Alpha 32
Keep n tokens 1
Max Token Length 150
Use xformers
設定を調整している部分はこれだけです。
学習に時間は掛かりますが自動で学習値を最適にしてくれます。
（高ステップで低学習率の方が高品質になると思っています）
最初の学習は教師画像４枚から始めるのが良いと思います。
衣装が統一された、close up、upper body、cowboy shot、full body、です。
出来上がったLORAを見て教師画像やステップを調整する。
I'm assuming you're using DreamboothLOLA with kohyaSS.
It is relatively easy if you have a high-quality teacher image of the character you want to make.
Train batch size, Epoch, adjusted to train each training image 750 to 1000 times
Mixed precision, Save precision, set according to the GPU you are using
LR Scheduler, Optimizer, Adafactor
Network Rank 64, Network Alpha 32
Keep n tokens 1
Max Token Length 150
Use xformers
That's all there is to adjusting the settings.
It takes time to learn, but it automatically optimizes the learning value.
(I think high steps and low learning rate will give higher quality)
I think that the first learning should start with 4 teacher images.
Close up, upper body, cowboy shot, full body, with unified costumes.
Adjust the teacher image and steps by looking at the finished LORA.
@tickling これで解りますか？
Do you understand this?
@jibunsagasinotabi Oh, I'm sorry.
I don't understand your detailed and appreciated explanation.
I'm a beginner so I don't really understand...what did you learn that from?
@tickling I thought you were asking me how to make it, but maybe I was wrong?
Here is the site for reference
作る方法を聞いているかと思ったが違ったのかな？
参考にしたサイトはここ
Maybe if you want them to make one, you won't be able to make a decision because you won't know if the material is available unless you write the character's name.
もしかしたら、"作って欲しい"なら、キャラクター名を書かないと、素材があるかも判らないから判断出来ない。