-
Notifications
You must be signed in to change notification settings - Fork 1
/
vis.py
44 lines (34 loc) · 1.68 KB
/
vis.py
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
import argparse
import glob
import os
import shutil
import torch
from PIL import Image
from torchvision.transforms import ToPILImage
from tqdm import tqdm
from model import Generator
from utils import get_transform
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='Vis Generator')
parser.add_argument('--sketch_name', default='/data/sketchy/val/sketch/cow', type=str,
help='Sketch image name')
parser.add_argument('--generator_name', default='result/sketchy_resnet50_512_generator.pth', type=str,
help='Generator name')
parser.add_argument('--save_root', default='result', type=str, help='Result saved root path')
opt = parser.parse_args()
sketch_names, generator_name, save_root = opt.sketch_name, opt.generator_name, opt.save_root
generator = Generator(in_channels=8, num_block=8)
generator.load_state_dict(torch.load(generator_name, map_location='cpu'))
generator = generator.cuda()
generator.eval()
sketch_names = glob.glob('{}/*.jpg'.format(sketch_names))
for sketch_name in tqdm(sketch_names):
sketch = get_transform('val')(Image.open(sketch_name)).unsqueeze(dim=0).cuda()
with torch.no_grad():
photo = generator(sketch)
result_path = '{}/{}'.format(save_root, os.path.basename(sketch_name).split('.')[0])
if os.path.exists(result_path):
shutil.rmtree(result_path)
os.mkdir(result_path)
Image.open(sketch_name).resize((224, 224), resample=Image.BILINEAR).save('{}/sketch.jpg'.format(result_path))
ToPILImage()((((photo.squeeze(dim=0) + 1.0) / 2) * 255).byte().cpu()).save('{}/photo.jpg'.format(result_path))