Hugging Face's logo

IDEA-CCNL
/
Taiyi-Stable-Diffusion-1B-Chinese-v0.1

Text-to-Image
Diffusers
Chinese
StableDiffusionPipeline
stable-diffusion
stable-diffusion-diffusers
Chinese
Inference Endpoints
Model card Files Community
37
weifeng-chen commited on
Commit
a398e36
•
1 Parent(s): 3cd0fc7

add image

Browse files
Files changed (1) hide show
  1. README.md +9 -8
README.md CHANGED
@@ -47,18 +47,19 @@ The first open source Chinese Stable diffusion, which was trained on 20M filtere
47
  We use [Noah-Wukong](https://wukong-dataset.github.io/wukong-dataset/)(100M) 和 [Zero](https://zero.so.com/)(23M) as our dataset, and take the image and text pairs with CLIP Score (based on [IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese](https://huggingface.co/IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese)) greater than 0.2 as our Training set. We use [IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese](https://huggingface.co/IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese) as our init text encoder. To keep the powerful generative capability of stable diffusion and align Chinese concepts with the images, We only train the text encoder and freeze other part of the [stable-diffusion-v1-4](https://huggingface.co/CompVis/stable-diffusion-v1-4)([paper](https://arxiv.org/abs/2112.10752)) model. It takes 100 hours to train this model based on 32 x A100. This model is a preliminary version and we will update this model continuously and open sourse. Welcome to exchange！
48
 
49
  ### Result
 
50
 
51
- 飞流直下三千尺，摄影。
52
- ![](result_examples/feiliu1.png)
 
53
 
54
- 飞流直下三千尺，油画。
55
- ![](result_examples/feiliu2.png)
 
 
 
56
 
57
- 铁马冰河入梦来，3D绘画。
58
- ![](result_examples/tiema1.png)
59
 
60
- 铁马冰河入梦来，概念画。
61
- ![](result_examples/tiema2.png)
62
  ## 使用 Usage
63
 
64
  ### 全精度 Full precision
 
47
  We use [Noah-Wukong](https://wukong-dataset.github.io/wukong-dataset/)(100M) 和 [Zero](https://zero.so.com/)(23M) as our dataset, and take the image and text pairs with CLIP Score (based on [IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese](https://huggingface.co/IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese)) greater than 0.2 as our Training set. We use [IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese](https://huggingface.co/IDEA-CCNL/Taiyi-CLIP-RoBERTa-102M-ViT-L-Chinese) as our init text encoder. To keep the powerful generative capability of stable diffusion and align Chinese concepts with the images, We only train the text encoder and freeze other part of the [stable-diffusion-v1-4](https://huggingface.co/CompVis/stable-diffusion-v1-4)([paper](https://arxiv.org/abs/2112.10752)) model. It takes 100 hours to train this model based on 32 x A100. This model is a preliminary version and we will update this model continuously and open sourse. Welcome to exchange！
48
 
49
  ### Result
50
+ Basic Prompt
51
 
52
+ | 铁马冰河入梦来，3D绘画。 | 飞流直下三千尺，油画。 | 女孩背影，日落，唯美插画。 |
53
+ | ---- | ---- | ---- |
54
+ | ![](result_examples/tiema.png) | ![](result_examples/feiliu.png) | ![](result_examples/nvhai.jpg) |
55
 
56
+ Advanced Prompt
57
+
58
+ | 铁马冰河入梦来，概念画，科幻，玄幻，3D | 中国海边城市，科幻，未来感，唯美，插画。 | 那人却在灯火阑珊处，色彩艳丽，古风，资深插画师作品，桌面高清壁纸。 |
59
+ | ---- | ---- | ---- |
60
+ | ![](result_examples/tiema2.jpg) | ![](result_examples/chengshi.jpg) | ![](result_examples/naren.jpg) |
61
 
 
 
62
 
 
 
63
  ## 使用 Usage
64
 
65
  ### 全精度 Full precision