Update README.md
Browse filesAdded "How to Use"
README.md
CHANGED
@@ -4,3 +4,54 @@ tags:
|
|
4 |
- PyTorch
|
5 |
- huggan
|
6 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
- PyTorch
|
5 |
- huggan
|
6 |
---
|
7 |
+
|
8 |
+
## How To Use
|
9 |
+
|
10 |
+
```python
|
11 |
+
|
12 |
+
from huggingface_hub import hf_hub_download
|
13 |
+
import torch
|
14 |
+
import matplotlib.pyplot as plt
|
15 |
+
import numpy as np
|
16 |
+
from torch import nn
|
17 |
+
|
18 |
+
class Generator(nn.Module):
|
19 |
+
def __init__(self):
|
20 |
+
super(Generator, self).__init__()
|
21 |
+
self.main = nn.Sequential(
|
22 |
+
nn.ConvTranspose2d(100, 64 * 8, 4, 1, 0, bias=False),
|
23 |
+
nn.BatchNorm2d(64 * 8),
|
24 |
+
nn.ReLU(True),
|
25 |
+
nn.ConvTranspose2d(64 * 8, 64 * 4, 4, 2, 1, bias=False),
|
26 |
+
nn.BatchNorm2d(64 * 4),
|
27 |
+
nn.ReLU(True),
|
28 |
+
nn.ConvTranspose2d(64 * 4, 64 * 2, 4, 2, 1, bias=False),
|
29 |
+
nn.BatchNorm2d(64 * 2),
|
30 |
+
nn.ReLU(True),
|
31 |
+
nn.ConvTranspose2d(64 * 2, 64, 4, 2, 1, bias=False),
|
32 |
+
nn.BatchNorm2d(64),
|
33 |
+
nn.ReLU(True),
|
34 |
+
nn.ConvTranspose2d(64, 3, 4, 2, 1, bias=False),
|
35 |
+
nn.Tanh()
|
36 |
+
)
|
37 |
+
|
38 |
+
def forward(self, input):
|
39 |
+
return self.main(input)
|
40 |
+
|
41 |
+
path = hf_hub_download('huggan/ArtGAN', 'ArtGAN.pt')
|
42 |
+
model = torch.load(path, map_location=torch.device('cpu'))
|
43 |
+
device = 'cuda' if torch.cuda.is_available() else 'cpu'
|
44 |
+
|
45 |
+
def generate(seed):
|
46 |
+
with torch.no_grad():
|
47 |
+
noise = torch.randn(seed, 100, 1, 1, device=device)
|
48 |
+
with torch.no_grad():
|
49 |
+
art = model(noise).detach().cpu()
|
50 |
+
gen = np.transpose(art[-1], (1, 2, 0))
|
51 |
+
fig = plt.figure(figsize=(5, 5))
|
52 |
+
plt.imshow(gen)
|
53 |
+
plt.axis('off')
|
54 |
+
|
55 |
+
generate(25)
|
56 |
+
|
57 |
+
```
|