/
generate.py
234 lines (185 loc) · 8.54 KB
/
generate.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
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
import argparse
import random
from matplotlib.image import imsave
import torch
import torch.nn as nn
from torchvision import utils
from model import Generator, Projection_module, Projection_module_church
from tqdm import tqdm
import sys
import os
def generate_gif(args, g_source, g_target, Proj_module):
if args.load_noise:
noise = torch.load(args.load_noise).cuda()
else:
noise = torch.randn(args.n_sample, args.latent).cuda()
with torch.no_grad():
n_steps = args.n_steps
step = float(1)/n_steps
n_paths = noise.size(0)
for t in range(n_paths):
print(t)
if t != (n_paths - 1):
z1, z2 = torch.unsqueeze(
noise[t], 0), torch.unsqueeze(noise[t+1], 0)
else:
z1, z2 = torch.unsqueeze(
noise[t], 0), torch.unsqueeze(noise[0], 0)
for i in range(n_steps):
alpha = step*i
z = z2*alpha + (1-alpha)*z1
sample_s, _ = g_source([z], randomize_noise=False)
w = [g_target.module.style(z)]
w = [Proj_module.modulate(item) for item in w]
sample_t, _= g_target(w, input_is_latent=True, randomize_noise=False)
utils.save_image(
sample_s,
f'%s/sample%d.png' % (args.save_source, (t*n_steps) + i) ,
normalize=True,
range=(-1, 1),
)
utils.save_image(
sample_t,
f'%s/sample%d.png' % (args.save_target, (t*n_steps) + i),
normalize=True,
range=(-1, 1),
)
def generate_imgs(args, g_source, g_target, Proj_module):
with torch.no_grad():
if args.load_noise:
sample_z = torch.load(args.load_noise)
else:
sample_z = torch.randn(args.n_sample, args.latent).cuda()
sample_s, _ = g_source([sample_z], input_is_latent=False, randomize_noise=False)
w = [g_target.module.style(sample_z)]
w = [Proj_module.modulate(item) for item in w]
sample_t, _= g_target(w, input_is_latent=True, randomize_noise=False)
utils.save_image(
sample_s,
f'%s/sample_s.png' % args.save_source,
nrow=5,
normalize=True,
range=(-1, 1),
)
utils.save_image(
sample_t,
f'%s/sample_t.png' % args.save_target,
nrow=5,
normalize=True,
range=(-1, 1),
)
def generate_img_pairs(args, g_source, g_target, Proj_module):
with torch.no_grad():
sample_z = torch.randn(args.SCS_samples, args.latent).cuda()
for i in range(10):
print(i)
w = [g_target.module.style(sample_z[i* int(args.SCS_samples / 10): (i+1)*int(args.SCS_samples / 10)])]
w = [Proj_module.modulate(item) for item in w]
sample_t, _= g_target(w, input_is_latent=True, randomize_noise=False)
sample_s, _ = g_source([sample_z[i* int(args.SCS_samples / 10): (i+1)*int(args.SCS_samples / 10)]], input_is_latent=False, randomize_noise=False)
for (num, (img_s, img_t)) in enumerate(zip(sample_s, sample_t)):
utils.save_image(
img_s,
f'%s/img%d.png' % (args.save_source, (i* int(args.SCS_samples / 10)) + num) ,
normalize=True,
range=(-1, 1),
)
utils.save_image(
img_t,
f'%s/img%d.png' % (args.save_target, (i * int(args.SCS_samples / 10)) + num) ,
normalize=True,
range=(-1, 1),
)
def generate_imgs_4IS(args, g_target, Proj_module):
with torch.no_grad():
sample_z = torch.randn(args.IS_sample, args.latent).cuda()
step = int(args.IS_sample / 50)
batch = 50
for i in range(int(args.IS_sample / 50)):
print(i)
w = [g_target.module.style(sample_z[i*batch: (i+1)*batch])]
w = [Proj_module.modulate(item) for item in w]
sample_t, _= g_target(w, input_is_latent=True, randomize_noise=False)
for (num, img) in enumerate(sample_t):
utils.save_image(
img,
f'%s/img%d.png' % (args.save_target, (i * batch) + num) ,
normalize=True,
range=(-1, 1),
)
if __name__ == '__main__':
device = 'cuda'
parser = argparse.ArgumentParser()
parser.add_argument('--size', type=int, default=256)
parser.add_argument('--SCS_samples', type=int, default=500, help='number of image pairs to eval SCS')
parser.add_argument('--n_sample', type=int, default=25, help='number of fake images to be sampled')
parser.add_argument('--IS_sample', type=int, default=10000, help='number of fake images to be sampled for IS')
parser.add_argument('--n_steps', type=int, default=40, help="determines the granualarity of interpolation")
parser.add_argument('--ckpt_source', type=str, default=None)
parser.add_argument('--ckpt_target', type=str, default=None)
parser.add_argument('--mode', type=str, default='viz_imgs', help='viz_imgs,viz_gif,eval_IS,eval_SCS')
parser.add_argument('--load_noise', type=str, default=None)
parser.add_argument('--channel_multiplier', type=int, default=2)
parser.add_argument('--target', type=str, default='VanGogh', help='target domain')
parser.add_argument('--task', type=int, default=10)
parser.add_argument('--source', type=str, default='face', help='source domain')
parser.add_argument('--latent_dir', type=str)
torch.manual_seed(10)
random.seed(10)
args = parser.parse_args()
args.latent = 512
args.n_mlp = 8
args.exp_name = args.target
if args.source == 'church':
Proj_module = Projection_module_church(args)
if args.source == 'face':
Proj_module = Projection_module(args)
if args.mode == 'viz_imgs' or args.mode == 'eval_SCS':
temp_str = f"%s2%s_%s" % (args.source, args.target, str(args.task))
imsave_path_source = os.path.join(args.mode, temp_str, 'source')
imsave_path_target = os.path.join(args.mode, temp_str, 'target')
if not os.path.exists(imsave_path_source):
os.makedirs(imsave_path_source)
if not os.path.exists(imsave_path_target):
os.makedirs(imsave_path_target)
args.save_source = imsave_path_source
args.save_target = imsave_path_target
if args.mode == 'viz_gif':
temp_str = f"%s_%s" % (args.source, str(args.task))
imsave_path_source = os.path.join(args.mode, temp_str, 'source')
imsave_path_target = os.path.join(args.mode, temp_str, args.target)
if not os.path.exists(imsave_path_source):
os.makedirs(imsave_path_source)
if not os.path.exists(imsave_path_target):
os.makedirs(imsave_path_target)
args.save_source = imsave_path_source
args.save_target = imsave_path_target
if args.mode == 'eval_IS':
temp_str = f"%s2%s_%s" % (args.source, args.target, str(args.task))
imsave_path_target = os.path.join(args.mode, temp_str)
if not os.path.exists(imsave_path_target):
os.makedirs(imsave_path_target)
args.save_target = imsave_path_target
# loading source model if available
if args.ckpt_source is not None:
g_source = Generator(
args.size, args.latent, args.n_mlp, channel_multiplier=args.channel_multiplier
).to(device)
checkpoint = torch.load(args.ckpt_source)
g_source.load_state_dict(checkpoint['g_ema'], strict=False)
# loading target model if available
if args.ckpt_target is not None:
g_target = Generator(
args.size, args.latent, args.n_mlp, channel_multiplier=args.channel_multiplier
).to(device)
g_target = nn.parallel.DataParallel(g_target)
checkpoint = torch.load(args.ckpt_target)
g_target.load_state_dict(checkpoint['g_ema'], strict=False)
if args.mode == 'viz_imgs':
generate_imgs(args, g_source, g_target, Proj_module)
if args.mode == 'eval_IS':
generate_imgs_4IS(args, g_target, Proj_module)
if args.mode == 'eval_SCS':
generate_img_pairs(args, g_source, g_target, Proj_module)
elif args.mode == 'viz_gif':
generate_gif(args, g_source, g_target, Proj_module)