Update README.md
Browse files
README.md
CHANGED
@@ -1,10 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
|
2 |
-
|
|
|
3 |
|
4 |
- Original authors: Zhengyuan Yang, Jianfeng Wang, Zhe Gan, Linjie Li, Kevin Lin, Chenfei Wu, Nan Duan, Zicheng Liu, Ce Liu, Michael Zeng, Lijuan Wang
|
5 |
- Original github repo by authors: https://github.com/microsoft/ReCo
|
6 |
- Converted to Diffusers: Jaemin Cho
|
7 |
|
|
|
|
|
|
|
|
|
|
|
8 |
# Example Usage
|
9 |
|
10 |
```python
|
@@ -92,5 +107,4 @@ boxes = [[263.68, 294.912, 380.544, 392.832], [121.344, 265.216, 267.392, 401.92
|
|
92 |
prompt = create_reco_prompt(caption, phrases, boxes)
|
93 |
prompt
|
94 |
>>> 'A box contains six donuts with varying types of glazes and toppings. <|endoftext|> <bin514> <bin575> <bin743> <bin766> <|startoftext|> chocolate donut. <|endoftext|> <bin237> <bin517> <bin522> <bin784> <|startoftext|> dark vanilla donut. <|endoftext|> <bin763> <bin575> <bin988> <bin745> <|startoftext|> donut with sprinkles. <|endoftext|> <bin234> <bin281> <bin524> <bin527> <|startoftext|> donut with powdered sugar. <|endoftext|> <bin515> <bin259> <bin767> <bin514> <|startoftext|> pink donut. <|endoftext|> <bin753> <bin290> <bin957> <bin506> <|startoftext|> brown donut. <|endoftext|>'
|
95 |
-
```
|
96 |
-
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
tags:
|
5 |
+
- reco
|
6 |
+
- text-to-image
|
7 |
+
- layout-to-image
|
8 |
+
pipeline_tag: text-to-image
|
9 |
+
---
|
10 |
|
11 |
+
|
12 |
+
# Diffusers 🧨 port of [ReCo: Region-Controlled Text-to-Image Generation (CVPR 2023)](https://arxiv.org/abs/2211.15518)
|
13 |
|
14 |
- Original authors: Zhengyuan Yang, Jianfeng Wang, Zhe Gan, Linjie Li, Kevin Lin, Chenfei Wu, Nan Duan, Zicheng Liu, Ce Liu, Michael Zeng, Lijuan Wang
|
15 |
- Original github repo by authors: https://github.com/microsoft/ReCo
|
16 |
- Converted to Diffusers: Jaemin Cho
|
17 |
|
18 |
+
|
19 |
+
# COCO checkpoint
|
20 |
+
- original pytorch lightning checkpoint: https://unitab.blob.core.windows.net/data/reco/reco_coco_616.ckpt
|
21 |
+
- original configuration yaml: https://github.com/microsoft/ReCo/blob/main/configs/reco/v1-finetune_cocogit.yaml
|
22 |
+
|
23 |
# Example Usage
|
24 |
|
25 |
```python
|
|
|
107 |
prompt = create_reco_prompt(caption, phrases, boxes)
|
108 |
prompt
|
109 |
>>> 'A box contains six donuts with varying types of glazes and toppings. <|endoftext|> <bin514> <bin575> <bin743> <bin766> <|startoftext|> chocolate donut. <|endoftext|> <bin237> <bin517> <bin522> <bin784> <|startoftext|> dark vanilla donut. <|endoftext|> <bin763> <bin575> <bin988> <bin745> <|startoftext|> donut with sprinkles. <|endoftext|> <bin234> <bin281> <bin524> <bin527> <|startoftext|> donut with powdered sugar. <|endoftext|> <bin515> <bin259> <bin767> <bin514> <|startoftext|> pink donut. <|endoftext|> <bin753> <bin290> <bin957> <bin506> <|startoftext|> brown donut. <|endoftext|>'
|
110 |
+
```
|
|