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 (9)
Tested this lora and I think it need absolutely same checkpoint, negative embeddings even sampler suggested in this articles and raising weights of prompt or adding detailed description prompts breaks the quality of output image I guess
@difficult133133
I understand that there are problems.
Since the test was a strong learning test, many checkpoints could not be generated well and many prompts could not be used.
We have confirmed that simple prompts can be used at several checkpoints.
不具合が有る事は把握しています。
強力な学習をさせるテストだったので、多くのチェックポイントで上手く生成出来ず、多くのプロンプトも使用出来ません。
簡素なプロンプトであれば数種類のチェックポイントで使用出来ている事を確認出来ています。
@jibunsagasinotabi yes I found it really strong when I put 1 boy in prompt and then her real boy friend came out lol. I think the strong learning would be good cause we can reach to the character's original look with less effort but that would be possible if I control the type of prompts and their weights. only hard thing is that those process could be a little bit difficult in case
@difficult133133
We are making it while doing various tests, so there are good times and bad times.
Although the subject is different, please try this improved LORA as well.
色々なテストをしながら作っているので、良い時もあれば悪い時もあります。
題材は違いますが、より改善されたこちらのLORAも試してみて下さい。
I tested more and found we can also use with many other models either samplers and upscalers. I also tested for some nsfws, nude stuffs and found lots of unnecessary lines generated in body. I tried many things as much as i can think of to get rid of them but I couldn't lol. Do you know the reason?
I don't think it can do difficult things because it is overtrained.
Try reducing the number of prompts.
for example,
Positive prompt: nsfw,1girl,full body,
Negative prompt: EasyNegative,
Check whether unnecessary lines appear in this example.
過学習させている為、難しい事は出来ないと思います。
プロンプト数を減らしてみて下さい。
例えば、
ポジティブプロンプト：nsfw,1girl,full body,
ネガティブプロンプト：EasyNegative,
この例で不要な線が出るか確認して下さい。
If unnecessary lines do not appear, try increasing the number of actions and backgrounds little by little.
If a line appears, the model is probably not suitable, so please change the model.
不要な線が出ないなら、行動や背景を少しずつ増やしてみて下さい。
線が出るなら、そのモデルは適合していないと思うのでモデルを変更して下さい。
I have uploaded a sample image with minimal prompts, please check it.
最小のプロンプトでサンプル画像をアップしておきました、確認して下さい。
I confirmed that I was able to output a good image with "MicchonShikimori," instead of "1girl."
"1girl"じゃなくて"MicchonShikimori,"でも良好な画像を出力出来た事を確認しました。
Positive prompt: nsfw,MicchonShikimori,full body,
Negative prompt: EasyNegative,
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.