change README
Browse files1st Edition with no sample
README.md
CHANGED
|
@@ -1,5 +1,51 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: other
|
| 3 |
-
license_name: fair-ai-public-license-1.0-sd
|
| 4 |
-
license_link: LICENSE
|
| 5 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
license_name: fair-ai-public-license-1.0-sd
|
| 4 |
+
license_link: LICENSE
|
| 5 |
+
---
|
| 6 |
+
22th, March, 2025,
|
| 7 |
+
This is a NoobAI v1.1 Epsiron Predict base, semireal model.
|
| 8 |
+
|
| 9 |
+
esotericism of noobai
|
| 10 |
+
(EoN)
|
| 11 |
+
|
| 12 |
+
recommend setting.
|
| 13 |
+
prompt:
|
| 14 |
+
*** MUST CHANGE THIS LINE *** character name, costume, situation, pose, and so on..., BREAK
|
| 15 |
+
(realistic:1.5), masterpiece, best quality, reference photo, photorealistic, cosplay photo, highres, very awa, photo background, detailed face, detailed background,
|
| 16 |
+
|
| 17 |
+
negative prompt:
|
| 18 |
+
lowres, (worst quality:1.3), bad quality, normal quality, bad anatomy, illustration, anime, (3d:1.5), game, cg, oekaki, figure, BREAK
|
| 19 |
+
watermark, signature, logo, artist name, text, username, stamp, date, footer, header, title, subtitle, body writing, censored,
|
| 20 |
+
|
| 21 |
+
sampler: Euler A
|
| 22 |
+
steps: 40 or higher(may be 28)
|
| 23 |
+
|
| 24 |
+
----
|
| 25 |
+
|
| 26 |
+
prompt:
|
| 27 |
+
- (realistic:1.5) for realistic image.
|
| 28 |
+
- realistic for semireal or 3d. to create character.
|
| 29 |
+
|
| 30 |
+
The goal is to make NoobAI v1.1 epsiron pred model more realistic, and it is the result of trial and error in an attempt to specialize in generating natural-looking cosplay images.
|
| 31 |
+
|
| 32 |
+
I started making it for myself and was angry about my environment crashing so much that I don't have any notes on how I made it, but I do know that I'm using my own LoRA. I put my own LoRA (also based on NoobAIv1.1epsiron pred) on Hugging Face.
|
| 33 |
+
|
| 34 |
+
The merge with LoRA was done using a modified version of a1111
|
| 35 |
+
-> https://github.com/frsk9999/sdwebui_savetensors_aftercreateimg
|
| 36 |
+
I think you can probably do the same thing using comfyui.
|
| 37 |
+
|
| 38 |
+
----
|
| 39 |
+
日本語で。
|
| 40 |
+
noobaiの秘教
|
| 41 |
+
|
| 42 |
+
という名前を付けました。
|
| 43 |
+
|
| 44 |
+
推奨プロンプトは英語部分を参照してください。
|
| 45 |
+
|
| 46 |
+
目的はNoobAI v1.1 epsiron predのリアル化です。
|
| 47 |
+
環境落ちすぎてメモれていない上に落ちたことに腹を立ててメモらずにもう進めるって言って進めたせいでレシピがないです(特にログに残らないマージ方法をとっていたのもあって…紛失デス)。
|
| 48 |
+
コンセプトとしてはreal_model_Nに近いです。つまりキャラクタの現界を目的にしています。
|
| 49 |
+
リアルなモデルがいろいろ出ていますが、衣装の再現に難があるものが多く、一旦混ぜてはオリジナルのnoobaiから層を持ってきてということを繰り返しています。
|
| 50 |
+
自作の怪しいツールを使っている所為か、ぐちゃぐちゃになっているので途中のモデルをベースにLoRAを作成しようとしたらNANになったりしました。
|
| 51 |
+
とはいってもこのモデルも完璧なものではなく結構衣装の再現ができないことが多いので、少しずつですが混ぜる比率等を変えながら進化させていきたいと思っています。
|