-
Notifications
You must be signed in to change notification settings - Fork 19
/
Copy pathfid_evaluation.py
182 lines (148 loc) · 7.32 KB
/
fid_evaluation.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
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
"""
Contains code for logging approximate FID scores during training.
If you want to output ground-truth images from the training dataset, you can
run this file as a script.
"""
import os
import shutil
import torch
import copy
import argparse
from torch.utils.data import dataset
from torchvision.utils import save_image
from pytorch_fid import fid_score
# from kid_score import calculate_kid_given_paths
from tqdm import tqdm
import datasets
import curriculums
COLOR_MAP = {
0: [0, 0, 0],
1: [204, 0, 0],
2: [76, 153, 0],
3: [204, 204, 0],
4: [51, 51, 255],
5: [204, 0, 204],
6: [0, 255, 255],
7: [255, 204, 204],
8: [102, 51, 0],
9: [255, 0, 0],
10: [102, 204, 0],
11: [255, 255, 0],
12: [0, 0, 153],
13: [0, 0, 204],
14: [255, 51, 153],
15: [0, 204, 204],
16: [0, 51, 0],
17: [255, 153, 51],
18: [0, 204, 0]}
def mask2color(masks):
masks = torch.argmax(masks, dim=1).float()
sample_mask = torch.zeros((masks.shape[0], masks.shape[1], masks.shape[2], 3), dtype=torch.float)
for key in COLOR_MAP:
sample_mask[masks==key] = torch.tensor(COLOR_MAP[key], dtype=torch.float)
sample_mask = sample_mask.permute(0,3,1,2)
return sample_mask
def output_real_images(dataloader, num_imgs, real_dir):
img_counter = 0
batch_size = dataloader.batch_size
dataloader = iter(dataloader)
for i in range(num_imgs//batch_size):
real_imgs, _ = next(dataloader)
for img in real_imgs:
save_image(img, os.path.join(real_dir, f'{img_counter:0>5}.jpg'), normalize=True, range=(-1, 1))
img_counter += 1
def output_real_labels_semantic(dataloader, num_imgs, real_dir):
img_counter = 0
batch_size = dataloader.batch_size
dataloader = iter(dataloader)
for i in range(num_imgs//batch_size):
_, real_labels, _ = next(dataloader)
for img in real_labels:
img = mask2color(img[None])[0]
save_image(img, os.path.join(real_dir, f'{img_counter:0>5}.jpg'), normalize=True, range=(-1, 1))
img_counter += 1
def setup_evaluation(dataset_name, generated_dir, dataset_path, target_size=128, num_imgs=8000, **kwargs):
# Only make real images if they haven't been made yet
if kwargs.get('background_mask', False):
real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_black_back' + '_real_images_' + str(target_size))
else:
real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_real_images_' + str(target_size))
if not os.path.exists(real_dir):
os.makedirs(real_dir)
dataloader, CHANNELS = datasets.get_dataset(dataset_name, dataset_path=dataset_path, img_size=target_size, background_mask=kwargs.get('background_mask', False), return_label=False)
print('outputting real images...')
output_real_images(dataloader, num_imgs, real_dir)
print('...done')
if generated_dir is not None:
os.makedirs(generated_dir, exist_ok=True)
return real_dir
def output_images(generator, input_metadata, rank, world_size, output_dir, num_imgs=2048):
metadata = copy.deepcopy(input_metadata)
metadata['img_size'] = 128
metadata['batch_size'] = 4
metadata['h_stddev'] = metadata.get('h_stddev_eval', metadata['h_stddev'])
metadata['v_stddev'] = metadata.get('v_stddev_eval', metadata['v_stddev'])
metadata['sample_dist'] = metadata.get('sample_dist_eval', metadata['sample_dist'])
metadata['psi'] = 1
img_counter = rank
generator.eval()
img_counter = rank
if rank == 0: pbar = tqdm("generating images", total = num_imgs)
with torch.no_grad():
while img_counter < num_imgs:
z = torch.randn((metadata['batch_size'], generator.module.z_dim), device=generator.module.device)
generated_imgs, _, _ = generator.module.staged_forward(z, **metadata)
for img in generated_imgs:
if img.shape[0] != 3:
img = img[-3:]
save_image(img, os.path.join(output_dir, f'{img_counter:0>5}.jpg'), normalize=True, range=(-1, 1))
img_counter += world_size
if rank == 0: pbar.update(world_size)
if rank == 0: pbar.close()
def output_images_double(generator, input_metadata, rank, world_size, output_dir, num_imgs=2048):
metadata = copy.deepcopy(input_metadata)
metadata['img_size'] = 128
metadata['batch_size'] = 4
metadata['h_stddev'] = metadata.get('h_stddev_eval', metadata['h_stddev'])
metadata['v_stddev'] = metadata.get('v_stddev_eval', metadata['v_stddev'])
metadata['sample_dist'] = metadata.get('sample_dist_eval', metadata['sample_dist'])
metadata['psi'] = 1
img_counter = rank
generator.eval()
img_counter = rank
if rank == 0: pbar = tqdm("generating images", total = num_imgs)
with torch.no_grad():
while img_counter < num_imgs:
z_geo = torch.randn((metadata['batch_size'], generator.module.z_geo_dim), device=generator.module.device)
z_app = torch.randn((metadata['batch_size'], generator.module.z_app_dim), device=generator.module.device)
generated_imgs, _ = generator.module.staged_forward(z_geo, z_app, **metadata)
for img in generated_imgs:
if img.shape[0] != 3:
img = img[-3:]
save_image(img, os.path.join(output_dir, f'{img_counter:0>5}.jpg'), normalize=True, range=(-1, 1))
img_counter += world_size
if rank == 0: pbar.update(world_size)
if rank == 0: pbar.close()
def calculate_fid(dataset_name, generated_dir, background_mask=False, target_size=256, **kwargs):
if background_mask:
real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_black_back' + '_real_images_' + str(target_size))
else:
real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_real_images_' + str(target_size))
fid = fid_score.calculate_fid_given_paths([real_dir, generated_dir], 128, 'cuda', 2048)
torch.cuda.empty_cache()
return fid
# def calculate_kid(dataset_name, generated_dir, background_mask=False, target_size=256, **kwargs):
# if background_mask:
# real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_black_back' + '_real_images_' + str(target_size))
# else:
# real_dir = os.path.join('/apdcephfs/share_1330077/starksun/projects/pi-GAN/data', 'EvalImages', dataset_name + '_real_images_' + str(target_size))
# kid = calculate_kid_given_paths([real_dir, generated_dir], 128, 'cuda', 2048)
# torch.cuda.empty_cache()
# return kid
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--dataset', type=str, default='CelebA')
parser.add_argument('--img_size', type=int, default=128)
parser.add_argument('--num_imgs', type=int, default=8000)
opt = parser.parse_args()
real_images_dir = setup_evaluation(opt.dataset, None, target_size=opt.img_size, num_imgs=opt.num_imgs)