minaichan commited on
Upload ccde141/Cute-style-reinforcement-Model/README.md with huggingface_hub
Browse files
ccde141/Cute-style-reinforcement-Model/README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: creativeml-openrail-m
|
| 3 |
+
---
|
| 4 |
+
本模型有三个国家的语言介绍
|
| 5 |
+
|
| 6 |
+
中文,英语,日文
|
| 7 |
+
|
| 8 |
+
This model has language introductions for three countries
|
| 9 |
+
|
| 10 |
+
Chinese, English, Japanese
|
| 11 |
+
|
| 12 |
+
このモデルには3つの国の言語が紹介されています
|
| 13 |
+
|
| 14 |
+
中国語、英語、日本語
|
| 15 |
+
|
| 16 |
+
=======================================
|
| 17 |
+
|
| 18 |
+
中文:基于NovalAI的原模型使用了15万张图片进行了训练,希望为大家带来一个有关于可爱风格的模型.稳定性目前尚可,后续我将会继续训练更多其他的模型.
|
| 19 |
+
|
| 20 |
+
下面是我对于模型训练的初衷与其他模型的注意事项:
|
| 21 |
+
|
| 22 |
+
初衷:希望可以探索到绘画ai的极致潜力
|
| 23 |
+
|
| 24 |
+
模型注意事项:
|
| 25 |
+
|
| 26 |
+
1.使用Hires.fix可以给予模型更优秀的输出效果。
|
| 27 |
+
|
| 28 |
+
2.适配很多基于SD1.5的lora,并可以有十分优秀的混合效果。
|
| 29 |
+
|
| 30 |
+
————————————————————————————
|
| 31 |
+
|
| 32 |
+
English: The original model based on NovalAI was trained using 150000 images, hoping to bring a cute style model to everyone. The stability is currently acceptable, and I will continue to train more other models in the future
|
| 33 |
+
|
| 34 |
+
Here are my initial intentions for model training and considerations for other models:
|
| 35 |
+
|
| 36 |
+
Original intention: I hope to explore the ultimate potential of AI in painting
|
| 37 |
+
|
| 38 |
+
Model considerations:
|
| 39 |
+
|
| 40 |
+
Using Hire.fix can provide the model with better output performance.
|
| 41 |
+
|
| 42 |
+
2. It is compatible with many SD1.5 based Loras and can have excellent mixing effects.
|
| 43 |
+
|
| 44 |
+
————————————————————————————
|
| 45 |
+
|
| 46 |
+
日文:NovalAIに基づくオリジナルモデルは15万枚の画像を使用して訓練を行って、みんなにかわいいスタイルのモデルを持ってきてほしい。安定性は今のところまだいいので、これからもっと他のモデルを訓練していきます。
|
| 47 |
+
|
| 48 |
+
次に、モデルトレーニングの初心と他のモデルについての注意事項を示します。
|
| 49 |
+
|
| 50 |
+
初心:絵画aiの究極の潜在力を探求したい
|
| 51 |
+
|
| 52 |
+
モデルの考慮事項:
|
| 53 |
+
|
| 54 |
+
1.Hires.fixを使用すると、モデルにより優れた出力効果を与えることができます。
|
| 55 |
+
|
| 56 |
+
2.SD 1.5ベースのローラを多く配合し、優れた混合効果を得ることができます。
|
| 57 |
+
|
| 58 |
+
=======================================
|