Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import torch
|
3 |
+
from model import Generator
|
4 |
+
import torchvision.utils as vutils
|
5 |
+
import os
|
6 |
+
from math import log2
|
7 |
+
|
8 |
+
# Function to generate images
|
9 |
+
def generate_images():
|
10 |
+
Z_DIM = 256
|
11 |
+
IN_CHANNELS = 256
|
12 |
+
|
13 |
+
# Load pretrained generator weights
|
14 |
+
checkpoint = torch.load("generator.pth", map_location=torch.device('cpu'))
|
15 |
+
|
16 |
+
# Filter out optimizer-related keys
|
17 |
+
state_dict = checkpoint['state_dict']
|
18 |
+
|
19 |
+
# Load the filtered state dictionary into the model
|
20 |
+
generator = Generator(Z_DIM, IN_CHANNELS, img_channels=3)
|
21 |
+
generator.load_state_dict(state_dict)
|
22 |
+
generator.eval()
|
23 |
+
|
24 |
+
# Set output directory
|
25 |
+
output_dir = "generated_images"
|
26 |
+
os.makedirs(output_dir, exist_ok=True)
|
27 |
+
|
28 |
+
# Generate images
|
29 |
+
img_sizes = [256]
|
30 |
+
images = []
|
31 |
+
for img_size in img_sizes:
|
32 |
+
num_steps = int(log2(img_size / 4))
|
33 |
+
x = torch.randn((6, Z_DIM, 1, 1)) # Generate a batch of 6 images
|
34 |
+
with torch.no_grad():
|
35 |
+
z = generator(x, alpha=0.5, steps=num_steps)
|
36 |
+
|
37 |
+
# Normalize the generated images to the range [-1, 1]
|
38 |
+
z = (z + 1) / 2
|
39 |
+
|
40 |
+
assert z.shape == (6, 3, img_size, img_size)
|
41 |
+
|
42 |
+
# Append generated images to the list
|
43 |
+
for i in range(6):
|
44 |
+
images.append(z[i].detach())
|
45 |
+
|
46 |
+
return images
|
47 |
+
|
48 |
+
# Main function to create Streamlit web app
|
49 |
+
def main():
|
50 |
+
st.title('Image Generation with pro-gan 🤖')
|
51 |
+
st.write("Click the buttons below to generate images.")
|
52 |
+
st.write("Trained on CelebHQ dataset.")
|
53 |
+
|
54 |
+
# Prompt message about image size
|
55 |
+
st.write("Note: Due to limited resources, the model has been trained to generate 256x256 size images. They are still awesome!")
|
56 |
+
|
57 |
+
# Generate images on button click
|
58 |
+
if st.button('Generate Images'):
|
59 |
+
images = generate_images()
|
60 |
+
# Display the generated images
|
61 |
+
for i, image in enumerate(images):
|
62 |
+
st.image(image.permute(1, 2, 0).cpu().numpy(), caption=f'Generated Image {i+1}', use_column_width=True)
|
63 |
+
|
64 |
+
if __name__ == '__main__':
|
65 |
+
main()
|