Spaces:
Configuration error
Configuration error
Update README.md
Browse files
README.md
CHANGED
@@ -1,17 +1,41 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
- MODELS: IP-Adapter, SDv.15: SG161222/Realistic_Vision_V4.0_noVAE, mattmdjaga/segformer_b2_clothes, lllyasviel/control_v11p_sd15_openpose
|
14 |
- If you have a high resolution outfit image, it is recommended to set OPENAI_CONSISTENCY_VAE = ENABLE in the .env file.
|
15 |
- If you want to use SG161222/Realistic_Vision_V5.1_noVAE, it is recommended that you do not change the default vae.
|
16 |
-
|
17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# OutfitChanger: Dress in any outfit you want.
|
2 |
+
PS : It is not currently possible to wear both bottoms and tops at same time.
|
3 |
+
## For deploy the model on Google Colab :
|
4 |
+
- !git clone https://huggingface.co/spaces/altayavci/OutfitChanger/
|
5 |
+
(PS: if you want clone the repo from Github please use <https://git-lfs.com>)
|
6 |
+
- cd /content/OutfitChanger
|
7 |
+
- !pip install -q gradio
|
8 |
+
- !pip install -r requirements.txt
|
9 |
+
- !python3 app.py
|
10 |
+
|
11 |
+
## CONFIGS:
|
12 |
+
|
13 |
+
| Library | Version |
|
14 |
+
|---------------------|------------|
|
15 |
+
| ip-adapter | https://github.com/tencent-ailab/IP-Adapter |
|
16 |
+
| torch | 2.1.0+cu121 |
|
17 |
+
| numpy | 1.23.5 |
|
18 |
+
| diffusers | 0.25.0 |
|
19 |
+
| transformers | 4.35.2 |
|
20 |
+
| accelerate | 0.25.0 |
|
21 |
+
| scipy | 1.11.4 |
|
22 |
+
| safetensors | 0.4.1 |
|
23 |
+
| controlnet_aux | 0.0.7 |
|
24 |
+
| pillow | 9.4.0 |
|
25 |
+
| opencv-python | 4.8.0 |
|
26 |
+
| Gradio | 4.13.0 |
|
27 |
+
|
28 |
+
|
29 |
+
|
30 |
- MODELS: IP-Adapter, SDv.15: SG161222/Realistic_Vision_V4.0_noVAE, mattmdjaga/segformer_b2_clothes, lllyasviel/control_v11p_sd15_openpose
|
31 |
- If you have a high resolution outfit image, it is recommended to set OPENAI_CONSISTENCY_VAE = ENABLE in the .env file.
|
32 |
- If you want to use SG161222/Realistic_Vision_V5.1_noVAE, it is recommended that you do not change the default vae.
|
33 |
+
|
34 |
+
* Django flask fastapi ile api yazıp docker kaldırılmalı
|
35 |
+
* İmg2txt’in verdiği cevaba göre seedler manipüle edilmeli
|
36 |
+
* Segmentasyon için sam ya da owlvit kullanılmalı
|
37 |
+
* Brush ile birnevi kullanıcıdan mask alınmalı
|
38 |
+
* Sd15 yerine SDXL kullanılmalı
|
39 |
+
* Referans kıyafetlerdeki sürekliliği artırmak için consistencydecoder kullanılabilir
|
40 |
+
* Eller ve ayaklara biraz daha focuslanılmalı
|
41 |
+
|