-
Notifications
You must be signed in to change notification settings - Fork 13
/
Copy pathmain.py
249 lines (188 loc) · 8.81 KB
/
main.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
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
import torch
import argparse
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
from os.path import join
from tensorboardX import SummaryWriter
from torchvision.utils import save_image
from collections import OrderedDict as OD
from torchvision import datasets, transforms, utils
from layers import IAFLayer
from utils import *
# Model definition
# ----------------------------------------------------------------------------------------------
class VAE(nn.Module):
def __init__(self, args):
super(VAE, self).__init__()
self.register_parameter('h', torch.nn.Parameter(torch.zeros(args.h_size)))
self.register_parameter('dec_log_stdv', torch.nn.Parameter(torch.Tensor([0.])))
layers = []
# build network
for i in range(args.depth):
layer = []
for j in range(args.n_blocks):
downsample = (i > 0) and (j == 0)
layer += [IAFLayer(args, downsample)]
layers += [nn.ModuleList(layer)]
self.layers = nn.ModuleList(layers)
self.first_conv = nn.Conv2d(3, args.h_size, 4, 2, 1)
self.last_conv = nn.ConvTranspose2d(args.h_size, 3, 4, 2, 1)
def forward(self, input):
# assumes input is \in [-0.5, 0.5]
x = self.first_conv(input)
kl, kl_obj = 0., 0.
h = self.h.view(1, -1, 1, 1)
for layer in self.layers:
for sub_layer in layer:
x = sub_layer.up(x)
h = h.expand_as(x)
self.hid_shape = x[0].size()
for layer in reversed(self.layers):
for sub_layer in reversed(layer):
h, curr_kl, curr_kl_obj = sub_layer.down(h)
kl += curr_kl
kl_obj += curr_kl_obj
x = F.elu(h)
x = self.last_conv(x)
x = x.clamp(min=-0.5 + 1. / 512., max=0.5 - 1. / 512.)
return x, kl, kl_obj
def sample(self, n_samples=64):
h = self.h.view(1, -1, 1, 1)
h = h.expand((n_samples, *self.hid_shape))
for layer in reversed(self.layers):
for sub_layer in reversed(layer):
h, _, _ = sub_layer.down(h, sample=True)
x = F.elu(h)
x = self.last_conv(x)
return x.clamp(min=-0.5 + 1. / 512., max=0.5 - 1. / 512.)
def cond_sample(self, input):
# assumes input is \in [-0.5, 0.5]
x = self.first_conv(input)
kl, kl_obj = 0., 0.
h = self.h.view(1, -1, 1, 1)
for layer in self.layers:
for sub_layer in layer:
x = sub_layer.up(x)
h = h.expand_as(x)
self.hid_shape = x[0].size()
outs = []
current = 0
for i, layer in enumerate(reversed(self.layers)):
for j, sub_layer in enumerate(reversed(layer)):
h, curr_kl, curr_kl_obj = sub_layer.down(h)
h_copy = h
again = 0
# now, sample the rest of the way:
for layer_ in reversed(self.layers):
for sub_layer_ in reversed(layer_):
if again > current:
h_copy, _, _ = sub_layer_.down(h_copy, sample=True)
again += 1
x = F.elu(h_copy)
x = self.last_conv(x)
x = x.clamp(min=-0.5 + 1. / 512., max=0.5 - 1. / 512.)
outs += [x]
current += 1
return outs
# Main
# ----------------------------------------------------------------------------------------------
if __name__ == '__main__':
# arguments
parser = argparse.ArgumentParser()
parser.add_argument('--debug', action='store_true')
parser.add_argument('--n_blocks', type=int, default=4)
parser.add_argument('--depth', type=int, default=2)
parser.add_argument('--z_size', type=int, default=32)
parser.add_argument('--h_size', type=int, default=64)
parser.add_argument('--n_epochs', type=int, default=1000)
parser.add_argument('--warmup', type=int, default=10)
parser.add_argument('--batch_size', type=int, default=32)
parser.add_argument('--free_bits', type=float, default=0.1)
parser.add_argument('--iaf', type=int, default=1)
parser.add_argument('--lr', type=float, default=1e-3)
args = parser.parse_args()
# create model and ship to GPU
model = VAE(args).cuda()
print(model)
# reproducibility is da best
set_seed(0)
opt = torch.optim.Adamax(model.parameters(), lr=args.lr)
# create datasets / dataloaders
scale_inv = lambda x : x + 0.5
ds_transforms = transforms.Compose([transforms.ToTensor(), lambda x : x - 0.5])
kwargs = {'num_workers':1, 'pin_memory':True, 'drop_last':True}
train_loader = torch.utils.data.DataLoader(datasets.CIFAR10('../cl-pytorch/data', train=True,
download=True, transform=ds_transforms), batch_size=args.batch_size, shuffle=True, **kwargs)
test_loader = torch.utils.data.DataLoader(datasets.CIFAR10('../cl-pytorch/data', train=False,
download=True, transform=ds_transforms), batch_size=args.batch_size, shuffle=True, **kwargs)
# spawn writer
model_name = 'NB{}_D{}_Z{}_H{}_BS{}_FB{}_LR{}_IAF{}'.format(args.n_blocks, args.depth, args.z_size, args.h_size,
args.batch_size, args.free_bits, args.lr, args.iaf)
model_name = 'test' if args.debug else model_name
log_dir = join('runs', model_name)
sample_dir = join(log_dir, 'samples')
writer = SummaryWriter(log_dir=log_dir)
maybe_create_dir(sample_dir)
print_and_save_args(args, log_dir)
print('logging into %s' % log_dir)
maybe_create_dir(sample_dir)
best_test = float('inf')
print('starting training')
for epoch in range(args.n_epochs):
model.train()
train_log = reset_log()
for batch_idx, (input,_) in enumerate(train_loader):
input = input.cuda()
x, kl, kl_obj = model(input)
log_pxz = logistic_ll(x, model.dec_log_stdv, sample=input)
loss = (kl_obj - log_pxz).sum() / x.size(0)
elbo = (kl - log_pxz)
bpd = elbo / (32 * 32 * 3 * np.log(2.))
opt.zero_grad()
loss.backward()
opt.step()
train_log['kl'] += [kl.mean()]
train_log['bpd'] += [bpd.mean()]
train_log['elbo'] += [elbo.mean()]
train_log['kl obj'] += [kl_obj.mean()]
train_log['log p(x|z)'] += [log_pxz.mean()]
for key, value in train_log.items():
print_and_log_scalar(writer, 'train/%s' % key, value, epoch)
print()
model.eval()
test_log = reset_log()
with torch.no_grad():
for batch_idx, (input,_) in enumerate(test_loader):
input = input.cuda()
x, kl, kl_obj = model(input)
log_pxz = logistic_ll(x, model.dec_log_stdv, sample=input)
loss = (kl_obj - log_pxz).sum() / x.size(0)
elbo = (kl - log_pxz)
bpd = elbo / (32 * 32 * 3 * np.log(2.))
test_log['kl'] += [kl.mean()]
test_log['bpd'] += [bpd.mean()]
test_log['elbo'] += [elbo.mean()]
test_log['kl obj'] += [kl_obj.mean()]
test_log['log p(x|z)'] += [log_pxz.mean()]
all_samples = model.cond_sample(input)
# save reconstructions
out = torch.stack((x, input)) # 2, bs, 3, 32, 32
out = out.transpose(1,0).contiguous() # bs, 2, 3, 32, 32
out = out.view(-1, x.size(-3), x.size(-2), x.size(-1))
all_samples += [x]
all_samples = torch.stack(all_samples) # L, bs, 3, 32, 32
all_samples = all_samples.transpose(1,0)
all_samples = all_samples.contiguous() # bs, L, 3, 32, 32
all_samples = all_samples.view(-1, x.size(-3), x.size(-2), x.size(-1))
save_image(scale_inv(all_samples), join(sample_dir, 'test_levels_{}.png'.format(epoch)), nrow=12)
save_image(scale_inv(out), join(sample_dir, 'test_recon_{}.png'.format(epoch)), nrow=12)
save_image(scale_inv(model.sample(64)), join(sample_dir, 'sample_{}.png'.format(epoch)), nrow=8)
for key, value in test_log.items():
print_and_log_scalar(writer, 'test/%s' % key, value, epoch)
print()
current_test = sum(test_log['bpd']) / batch_idx
if current_test < best_test:
best_test = current_test
print('saving best model')
torch.save(model.state_dict(), join(log_dir, 'best_model.pth'))