-
Notifications
You must be signed in to change notification settings - Fork 1
/
train.py
172 lines (148 loc) · 8.86 KB
/
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
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
import time, datetime
import os
import tensorflow as tf
import numpy as np
from cnn_model import SentenceCNN
import preprocessing
###########################################
# Defining Parameters
###########################################
tf.flags.DEFINE_integer("n_classes", 6, "Number of classes")
# Model Hyperparameters
tf.flags.DEFINE_integer("embedding_dim", 128, "Dimensionality of character embedding (default: 128)")
tf.flags.DEFINE_string("filter_sizes", "3,4,5", "Comma-separated filter sizes (default: '3,4,5')")
tf.flags.DEFINE_integer("n_filters_per_size", 100, "Number of filters per filter size (default: 128)")
tf.flags.DEFINE_float("dropout_keep_prob", 0.5, "Dropout keep probability (default: 0.5)")
tf.flags.DEFINE_float("lambda_l2_reg", 0.1, "L2 regularizaion lambda (default: 0.0)")
tf.flags.DEFINE_float("l2_norm_cutoff", 3.0, "Cutoff for l2 norm regularization of weights (default: 3.0)")
# Training parameters
tf.flags.DEFINE_integer("batch_size", 50, "Batch Size (default: 64)")
tf.flags.DEFINE_integer("n_epochs", 10, "Number of training epochs (default: 200)")
tf.flags.DEFINE_integer("evaluate_every", 100, "Evaluate model on dev set after this many steps (default: 100)")
tf.flags.DEFINE_integer("checkpoint_every", 100, "Save model after this many steps (default: 100)")
# Misc Parameters
tf.flags.DEFINE_boolean("allow_soft_placement", True, "Allow device soft device placement")
tf.flags.DEFINE_boolean("log_device_placement", False, "Log placement of ops on devices")
lambda_l2_param_search = [0.001, 0.01, 0.1, 1, 10]
embedding_dim_param_search = [64, 128, 256]
tf.flags.FLAGS._parse_flags()
print("\nParameters:")
for attr, value in sorted(tf.flags.FLAGS.__flags.items()):
print("{}={}".format(attr.upper(), value))
print("")
###########################################
# Data Preparation
###########################################
print("loading data...")
x, y, vocab, vocab_inv = preprocessing.load_data()
# Shuffle data randomly
np.random.seed(10)
# shuffled_indices = np.random.permutation(np.arange(len(y)))
# x_shuffled = x[shuffled_indices]
# y_shuffled = y[shuffled_indices]
# split train/valid set
# TODO Implement a fucking correct XVal procedure for this
x_train, x_valid = x[:-500], x[-500:]
y_train, y_valid = y[:-500], y[-500:]
print("Vocabulary Size: {:d}".format(len(vocab)))
print("Train/Validation split: {:d}/{:d}".format(len(y_train), len(y_valid)))
###########################################
# Training
###########################################
for lambda_l2 in lambda_l2_param_search:
for embed_dim in embedding_dim_param_search:
print("lambd_l2 = {:g}, embed_dim = {}".format(lambda_l2, embed_dim))
with tf.Graph().as_default():
session_conf = tf.ConfigProto(allow_soft_placement=tf.flags.FLAGS.allow_soft_placement,
log_device_placement=tf.flags.FLAGS.log_device_placement)
sess = tf.Session(config=session_conf)
with sess.as_default():
cnn = SentenceCNN(
seq_len=x_train.shape[1],
n_classes=tf.flags.FLAGS.n_classes,
vocab_size=len(vocab),
embedding_size=embed_dim,
filter_sizes=list(map(int, tf.flags.FLAGS.filter_sizes.split(","))),
n_filters_per_size=tf.flags.FLAGS.n_filters_per_size,
lambda_l2_reg=lambda_l2)
global_step = tf.Variable(0, name="global_step", trainable=False)
optimizer = tf.train.AdamOptimizer(1e-3)
grads_and_vars = optimizer.compute_gradients(cnn.loss)
train_op = optimizer.apply_gradients(grads_and_vars=grads_and_vars, global_step=global_step)
# here we keep track of gradient values and sparsity
grad_summaries = []
for g, v in grads_and_vars:
if g is not None:
grad_hist_summary = tf.histogram_summary("{}/grad/hist".format(v.name), g)
sparsity_summary = tf.scalar_summary("{}/grad/sparsity".format(v.name), tf.nn.zero_fraction(g))
grad_summaries.append(grad_hist_summary)
grad_summaries.append(sparsity_summary)
grad_summaries_merged = tf.merge_summary(grad_summaries)
# Output directory for models and summaries
timestamp = str(int(time.time()))
out_dir = os.path.abspath(os.path.join(os.path.curdir, "runs", timestamp))
print("Writing to {}\n".format(out_dir))
# Summaries for loss and accuracy
loss_summary = tf.scalar_summary("loss", cnn.loss)
acc_summary = tf.scalar_summary("accuracy", cnn.accuracy)
# Train Summaries
train_summary_op = tf.merge_summary([loss_summary, acc_summary, grad_summaries_merged])
train_summary_dir = os.path.join(out_dir, "summaries", "train")
train_summary_writer = tf.train.SummaryWriter(train_summary_dir, sess.graph)
# Dev summaries
dev_summary_op = tf.merge_summary([loss_summary, acc_summary])
dev_summary_dir = os.path.join(out_dir, "summaries", "dev")
dev_summary_writer = tf.train.SummaryWriter(dev_summary_dir, sess.graph)
# Checkpoint directory. Tensorflow assumes this directory already exists so we need to create it
checkpoint_dir = os.path.abspath(os.path.join(out_dir, "checkpoints"))
checkpoint_prefix = os.path.join(checkpoint_dir, "model")
if not os.path.exists(checkpoint_dir):
os.makedirs(checkpoint_dir)
saver = tf.train.Saver(tf.all_variables())
sess.run(tf.initialize_all_variables())
def train_step(x_batch, y_batch):
feed_dict = {cnn.x: x_batch, cnn.y: y_batch, cnn.dropout_keep_prob: tf.flags.FLAGS.dropout_keep_prob}
_, step, summaries, loss, accuracy = sess.run([train_op, global_step, train_summary_op,
cnn.loss, cnn.accuracy], feed_dict=feed_dict)
time_str = datetime.datetime.now().isoformat()
# print("{}: step {}, loss {:g}, acc {:g}".format(time_str, step, loss, accuracy))
train_summary_writer.add_summary(summaries, step)
# update the weight values after each training step
W_fc = cnn.W_fc.eval()
l2_norm = np.linalg.norm(W_fc)
# print(l2_norm)
if l2_norm > tf.flags.FLAGS.l2_norm_cutoff:
# first scale the matrix so the norm is 1 and then multiply by the cutoff
W_fc /= l2_norm
W_fc *= tf.flags.FLAGS.l2_norm_cutoff
assign_op = cnn.W_fc.assign(W_fc)
sess.run(assign_op)
W_fc = cnn.W_fc.eval()
l2_norm = np.linalg.norm(W_fc)
# print(">>>>>>> NEW L2 NORM IS:" + str(l2_norm))
def validation_step(x_batch, y_batch, writer=None):
feed_dict = {cnn.x: x_batch, cnn.y: y_batch, cnn.dropout_keep_prob: 1.0}
step, summaries, loss, accuracy = sess.run([global_step, dev_summary_op, cnn.loss, cnn.accuracy],
feed_dict=feed_dict)
time_str = datetime.datetime.now().isoformat()
print("{}: step {}, loss {:g}, acc {:g}".format(time_str, step, loss, accuracy))
if writer:
writer.add_summary(summaries, step)
# Generate batches
batches = preprocessing.batch_iter(list(zip(x_train, y_train)), tf.flags.FLAGS.batch_size,
tf.flags.FLAGS.n_epochs)
# Training loop. For each batch...
for batch in batches:
x_batch, y_batch = zip(*batch)
train_step(x_batch, y_batch)
current_step = tf.train.global_step(sess, global_step)
# if current_step % tf.flags.FLAGS.evaluate_every == 0:
# print("\nEvaluation on Validation set:")
# validation_step(x_valid, y_valid, writer=dev_summary_writer)
# print("")
# if current_step % tf.flags.FLAGS.checkpoint_every == 0:
# path = saver.save(sess, checkpoint_prefix, global_step=current_step)
# print("Saved model checkpoint to {}\n".format(path))
print("\nEvaluation on Validation set:")
validation_step(x_valid, y_valid, writer=dev_summary_writer)
print("")