/
vae.py
138 lines (124 loc) · 5.9 KB
/
vae.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
import numpy as np
from theano import function, shared
import theano.tensor as T
import cPickle as pk
from mlp import GaussianMLP, BernoulliMLP
from utils import kld_unit_mvn, kldu_unit_mvn, load_dataset, floatX
import time
ADAG_EPS = 1e-12 # for stability ##1e-10
class VAE(object):
def __init__(self, xdim, args, dec_nonlin=None):
self.xdim = xdim
self.hdim = args.hdim
self.zdim = args.zdim
self.lmbda = args.lmbda # weight decay coefficient * 2
self.x = T.matrix('x', dtype=floatX)
self.eps = T.matrix('eps', dtype=floatX)
self.train_i = T.scalar('train_i', dtype=floatX)
self.dec = args.decM
self.COV = args.COV
self.enc_mlp = GaussianMLP(self.x, self.xdim, self.hdim, self.zdim, nlayers=args.nlayers, eps=self.eps, COV=self.COV)
if self.dec == 'bernoulli':
# log p(x | z) defined as -CE(x, y) = dec_mlp.cost(y)
self.dec_mlp = BernoulliMLP(self.enc_mlp.out, self.zdim, self.hdim, self.xdim, nlayers=args.nlayers, y=self.x)
elif self.dec == 'gaussian':
self.dec_mlp = GaussianMLP(self.enc_mlp.out, self.zdim, self.hdim, self.xdim, nlayers=args.nlayers, y=self.x, activation=dec_nonlin, COV=self.COV)
else:
raise RuntimeError('unrecognized decoder %' % dec)
#encoder part + decoder part
if self.COV == False:
self.enc_cost = -T.sum(kld_unit_mvn(self.enc_mlp.mu, self.enc_mlp.var))
else:
self.enc_cost = -T.sum(kldu_unit_mvn(self.enc_mlp.mu, self.enc_mlp.var, self.enc_mlp.u))
self.cost = (self.enc_cost + self.dec_mlp.cost) / args.batsize
self.params = self.enc_mlp.params + self.dec_mlp.params
##[T.grad(self.cost, p) + self.lmbda * p for p in self.params]
self.gparams = [T.grad(self.cost, p) for p in self.params]
self.gaccums = [shared(value=np.zeros(p.get_value().shape, dtype=floatX)) for p in self.params]
self.lr = args.lr * (1-args.lmbda)**self.train_i
# update params, update sum(grad_params) for adagrade
self.updates = [
(param, param - self.lr*gparam/T.sqrt(gaccum+T.square(gparam)+ADAG_EPS))
for param, gparam, gaccum in zip(self.params, self.gparams, self.gaccums) ]
self.updates += [ (gaccum, gaccum + T.square(gparam))
for gaccum, gparam in zip(self.gaccums, self.gparams) ]
self.train = function(
inputs=[self.x, self.eps, self.train_i],
outputs=self.cost,
updates=self.updates
)
self.test = function(
inputs=[self.x, self.eps],
outputs=self.cost,
updates=None
)
# can be used for semi-supervised learning for example
self.encode = function(
inputs=[self.x, self.eps],
outputs=self.enc_mlp.out
)
# use this to sample
self.decode = function(
inputs=[self.enc_mlp.out], ##z with shape (1,2)
outputs=self.dec_mlp.out
) ##mlp103 .out=.mu+.sigma*eps
def main():
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-batsize', default=100) ##100
parser.add_argument('-nlayers', default=1, type=int, help='num_hid_layers before output')
parser.add_argument('-hdim', default=200, type=int) ##200 for freyfaces
parser.add_argument('-zdim', default=2, type=int) ##2
parser.add_argument('-lmbda', default=0., type=float, help='weight decay coeff') ##0.001
parser.add_argument('-lr', default=0.01, type=float, help='learning rate') ##0.01
parser.add_argument('-epochs', default=100, type=int) ##1000
parser.add_argument('-print_every', default=5, type=int) ##100
parser.add_argument('-save_every', default=50, type=int) ##1
parser.add_argument('-outfile', default='vae_model.pk')
parser.add_argument('-dset', default='mnist') ##mnist freyfaces
parser.add_argument('-COV', default=False, type=bool)
parser.add_argument('-decM', default='gaussian', help='bernoulli | gaussian')
args = parser.parse_args()
batsize = args.batsize
dset = args.dset
data = load_dataset(dset)
valid_fg = 0
dec_nonlin = T.nnet.relu ##T.nnet.softplus
if dset=='mnist':
train_x, train_y = data['train'] ##mnist: (N,784)
valid_x, valid_y = data['valid']
num_valid_bats = valid_x.shape[0] / batsize
print "valid data shape: ", valid_x.shape
valid_fg = 1
elif dset=='freyfaces':
train_x = data
print "training data shape: ", train_x.shape
model = VAE(train_x.shape[1], args, dec_nonlin=dec_nonlin)
num_train_bats = train_x.shape[0] / batsize ##discard last <batsize
begin = time.time()
for i in xrange(args.epochs):
for k in xrange(num_train_bats):
x = train_x[k*batsize : (k+1)*batsize, :]
eps = np.random.randn(x.shape[0], args.zdim).astype(floatX)
cost = model.train(x, eps, i) ##update_times=epochs*num_train_bats
j = i+1
if j % args.print_every == 0: ##(b+1)
end = time.time()
print('epoch %d, cost %.2f, time %.2fs' % (j, cost, end-begin))
begin = end
if valid_fg == 1:
valid_cost = 0
for l in xrange(num_valid_bats):
x_val = valid_x[l*batsize:(l+1)*batsize, :]
eps_val = np.zeros((x_val.shape[0], args.zdim), dtype=floatX)
valid_cost = valid_cost + model.test(x_val, eps_val)
valid_cost = valid_cost / num_valid_bats
print('valid cost: %f' % valid_cost)
if j % args.save_every == 0: ##
with open(args.outfile, 'wb') as f:
pk.dump(model, f, protocol=pk.HIGHEST_PROTOCOL)
print('model saved')
# with open(args.outfile, 'wb') as f:
# pk.dump(model, f, protocol=pk.HIGHEST_PROTOCOL)
if __name__ == '__main__':
main()