-
Notifications
You must be signed in to change notification settings - Fork 11
/
Copy pathosda_train.py
157 lines (132 loc) · 7.19 KB
/
osda_train.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
import os, sys
import numpy as np
import tensorflow as tf
import datetime
from model import DALearner
from utils import utils
from utils import metrics
from preprocessing.preprocessing import preprocessing
import math
tf.app.flags.DEFINE_float('lr', '1e-3', 'Learning rate for adam optimizer')
tf.app.flags.DEFINE_float('dropout_keep_prob', 0.5, 'Dropout keep probability')
tf.app.flags.DEFINE_integer('num_epochs', 200, 'Number of epochs for training')
tf.app.flags.DEFINE_integer('batch_size', 128, 'Batch size')
tf.app.flags.DEFINE_string('net','small', '[small,large]')
tf.app.flags.DEFINE_string('opt','mom', '[adam,mom]')
tf.app.flags.DEFINE_string('train','mnist', '[mnist,usps,svshn]')
tf.app.flags.DEFINE_string('test','usps', '[mnist,usps,svshn]')
tf.app.flags.DEFINE_string('train_root_dir', '../training', 'Root directory to put the training data')
tf.app.flags.DEFINE_integer('log_step', 10000, 'Logging period in terms of iteration')
#-------------------------open set domain adaptation----------------------------------------
NUM_CLASSES = 6
FLAGS = tf.app.flags.FLAGS
TRAIN_FILE=FLAGS.train
TEST_FILE=FLAGS.test
print TRAIN_FILE+' ---------------------------------------> '+TEST_FILE
print TRAIN_FILE+' ---------------------------------------> '+TEST_FILE
print TRAIN_FILE+' ---------------------------------------> '+TEST_FILE
TRAIN=utils.get_data(FLAGS.train,split='train',unk=False,shuffle=True,frange=[0.,1.])
VALID=utils.get_data(FLAGS.test,split='train',unk=True,shuffle=True,frange=[0.,1.])
TEST=utils.get_data(FLAGS.test,split='train',unk=True,shuffle=False,frange=[0.,1.])
def adaptation_factor(x):
den=1.0+math.exp(-10*x)
lamb=2.0/den-1.0
return min(lamb,1.0)
def main(_):
# Create training directories
now = datetime.datetime.now()
train_dir_name = now.strftime('alexnet_%Y%m%d_%H%M%S')
train_dir = os.path.join(FLAGS.train_root_dir, train_dir_name)
checkpoint_dir = os.path.join(train_dir, 'checkpoint')
tensorboard_dir = os.path.join(train_dir, 'tensorboard')
tensorboard_train_dir = os.path.join(tensorboard_dir, 'train')
tensorboard_val_dir = os.path.join(tensorboard_dir, 'val')
if not os.path.isdir(FLAGS.train_root_dir): os.mkdir(FLAGS.train_root_dir)
if not os.path.isdir(train_dir): os.mkdir(train_dir)
if not os.path.isdir(checkpoint_dir): os.mkdir(checkpoint_dir)
if not os.path.isdir(tensorboard_dir): os.mkdir(tensorboard_dir)
if not os.path.isdir(tensorboard_train_dir): os.mkdir(tensorboard_train_dir)
if not os.path.isdir(tensorboard_val_dir): os.mkdir(tensorboard_val_dir)
dropout_keep_prob = tf.placeholder(tf.float32)
revgrad_lamb = tf.placeholder(tf.float32)
is_training=tf.placeholder(tf.bool)
# Model
model =DALearner(name=FLAGS.net,num_classes=NUM_CLASSES,source=FLAGS.train,target=FLAGS.test)
# Placeholders
x_s = tf.placeholder(tf.float32, [None]+TRAIN.image_shape,name='x')
x_t = tf.placeholder(tf.float32, [None]+TEST.image_shape,name='xt')
x=preprocessing(x_s,model)
xt=preprocessing(x_t,model)
tf.summary.image('Source Images',x)
tf.summary.image('Target Images',xt)
y = tf.placeholder(tf.float32, [None, NUM_CLASSES],name='y')
yt = tf.placeholder(tf.float32, [None, NUM_CLASSES],name='yt')
loss_class,loss_gen,src_p,trg_p= model.loss(x, y, xt, keep_prob=dropout_keep_prob,phase=is_training,lamb=revgrad_lamb)
#---- Optimizers---------
main_vars=tf.trainable_variables()
gen_vars=[var for var in main_vars if 'gen' in var.name]
class_vars=[var for var in main_vars if 'class' in var.name]
print gen_vars
print class_vars
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
gen_op=utils.get_optimizer(FLAGS.opt,FLAGS.lr,loss_gen,gen_vars)
class_op=utils.get_optimizer(FLAGS.opt,FLAGS.lr,loss_class,class_vars)
optimizer=tf.group(gen_op,class_op)
#------------ A series of metrics for evaluation: OS,OS*,ALL,UNK----------------------------------
target_predict=trg_p
with tf.variable_scope('metrics') as scope:
os_acc,os_update_op=metrics.OS(hx=target_predict,y=yt,num_classes=NUM_CLASSES)
osstar_acc,osstar_update_op=metrics.OS_star(hx=target_predict,y=yt,num_classes=NUM_CLASSES)
all_acc,all_update_op=metrics.ALL(hx=target_predict,y=yt)
unk_acc,unk_update_op=metrics.UNK(hx=target_predict,y=yt,num_classes=NUM_CLASSES)
metrics_update_op=tf.group(os_update_op,osstar_update_op,all_update_op,unk_update_op)
metrics_variables=[v for v in tf.local_variables() if v.name.startswith('metrics')]
reset_ops=[v.initializer for v in metrics_variables]
print metrics_variables
train_writer=tf.summary.FileWriter('./log/tensorboard')
train_writer.add_graph(tf.get_default_graph())
merged=tf.summary.merge_all()
print '============================GLOBAL TRAINABLE VARIABLES ============================'
print tf.trainable_variables(),' ',len(tf.trainable_variables())
#print '============================GLOBAL VARIABLES ======================================'
#print tf.global_variables()
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
sess.run(tf.local_variables_initializer())
saver=tf.train.Saver()
train_writer.add_graph(sess.graph)
print("{} Start training...".format(datetime.datetime.now()))
for step in range(200*600):
# Start training
batch_xs, batch_ys = TRAIN.next_batch(FLAGS.batch_size)
Tbatch_xs, Tbatch_ys = VALID.next_batch(FLAGS.batch_size)
MAX_STEP=10000
constant=adaptation_factor(step*1.0/MAX_STEP)
summary,_=sess.run([merged,optimizer], feed_dict={x_s: batch_xs,x_t: Tbatch_xs,is_training:True,y: batch_ys,revgrad_lamb:constant,dropout_keep_prob:0.5,yt:Tbatch_ys})
train_writer.add_summary(summary,step)
if step%600==0:
epoch=step/600
print("{} Start validation".format(datetime.datetime.now()))
#print 'Epoch {0:<10} Step {1:<10} C_loss {2:<10} Advloss {3:<10}'.format(epoch,step,closs,advloss)
test_acc = 0.
test_count = 0
bs=500
print constant
print 'test_counts ',len(TEST.labels)
for _ in xrange((len(TEST.labels))/bs):
batch_tx, batch_ty = TEST.next_batch(bs)
sess.run(metrics_update_op, feed_dict={x_t: batch_tx, yt: batch_ty, is_training:False,dropout_keep_prob: 1.})
osacc,osstaracc,allacc,unkacc = sess.run([os_acc,osstar_acc,all_acc,unk_acc], feed_dict={x_t: batch_tx, yt: batch_ty, is_training:False,dropout_keep_prob: 1.})
test_count += bs
res=len(TEST.labels)%bs
if res>0:
batch_tx, batch_ty = TEST.next_batch(res)
sess.run(metrics_update_op, feed_dict={x_t: batch_tx, yt: batch_ty, is_training:False,dropout_keep_prob: 1.})
osacc,osstaracc,allacc,unkacc = sess.run([os_acc,osstar_acc,all_acc,unk_acc], feed_dict={x_t: batch_tx, yt: batch_ty, is_training:False,dropout_keep_prob: 1.})
print "Epoch {4:<5} OS {0:<10} OS* {1:<10} ALL {2:<10} UNK {3:<10}".format(osacc,osstaracc,allacc,unkacc,epoch)
sess.run(reset_ops)
if epoch==300:
return
if __name__ == '__main__':
tf.app.run()