File size: 1,686 Bytes
ac4ce84
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
# face-parsing.PyTorch

<p align="center">
	<a href="https://github.com/zllrunning/face-parsing.PyTorch">
    <img class="page-image" src="https://github.com/zllrunning/face-parsing.PyTorch/blob/master/6.jpg" >
	</a>
</p>

### Contents
- [Training](#training)
- [Demo](#Demo)
- [References](#references)

## Training

1. Prepare training data:
    -- download [CelebAMask-HQ dataset](https://github.com/switchablenorms/CelebAMask-HQ)

	--  change file path in the `prepropess_data.py`  and run
```Shell
python prepropess_data.py
```

2. Train the model using CelebAMask-HQ dataset:
Just run the train script: 
```
    $ CUDA_VISIBLE_DEVICES=0,1 python -m torch.distributed.launch --nproc_per_node=2 train.py
```

If you do not wish to train the model, you can download [our pre-trained model](https://drive.google.com/open?id=154JgKpzCPW82qINcVieuPH3fZ2e0P812) and save it in `res/cp`.


## Demo
1. Evaluate the trained model using:
```Shell
# evaluate using GPU
python test.py
```

## Face makeup using parsing maps
[**face-makeup.PyTorch**](https://github.com/zllrunning/face-makeup.PyTorch)
<table>

<tr>
<th>&nbsp;</th>
<th>Hair</th>
<th>Lip</th>
</tr>

<!-- Line 1: Original Input -->
<tr>
<td><em>Original Input</em></td>
<td><img src="makeup/116_ori.png" height="256" width="256" alt="Original Input"></td>
<td><img src="makeup/116_lip_ori.png" height="256" width="256" alt="Original Input"></td>
</tr>

<!-- Line 3: Color -->
<tr>
<td>Color</td>
<td><img src="makeup/116_1.png" height="256" width="256" alt="Color"></td>
<td><img src="makeup/116_3.png" height="256" width="256" alt="Color"></td>
</tr>

</table>


## References
- [BiSeNet](https://github.com/CoinCheung/BiSeNet)