text_convnet.py

import tensorflow as tf
import numpy as np

class TextConvNet(object):
	def __init__(self, sequence_length, num_classes, vocab_size,
				 embedding_size, filter_sizes, num_filters, l2_reg_lambda=0.0):
		# Placeholders for input vectors and dropout probability
		self.input_x = tf.placeholder(tf.int32, [None, sequence_length], name="input_x")
		self.input_y = tf.placeholder(tf.float32, [None, num_classes], name="input_y")
		self.dropout_keep_probability = tf.placeholder(tf.float32, name="dropout_keep_probability")

		# Track L2 Loss
		l2_loss = tf.constant(0.0)

		with tf.device('/cpu:0'), tf.name_scope("embedding"):
			W = tf.Variable(tf.random_uniform([vocab_size, embedding_size], -1.0, 1.0), name="W")
			self.embedded_chars = tf.nn.embedding_lookup(W, self.input_x)
			self.embedded_chars_expanded = tf.expand_dims(self.embedded_chars, -1)

		# Create a convolution layer and a pool for each filter size
		pooled_out = []
		for filter_size in filter_sizes:
			with tf.name_scope("conv-maxpool-%s" % filter_size):
				# Convolution layer
				filter_shape = [filter_size, embedding_size, 1, num_filters]
				W = tf.Variable(tf.truncated_normal(filter_shape, stddev=0.1), name="W")
				b = tf.Variable(tf.constant(0.1, shape=[num_filters]), name="b")
				conv = tf.nn.conv2d(self.embedded_chars_expanded,
									W,
									strides=[1, 1, 1, 1],
									padding="VALID",
									name="conv")
				# Apply non-linearity
				h = tf.nn.relu(tf.nn.bias_add(conv, b), name="relu")
				# Max Pooling
				pooled = tf.nn.max_pool(h,
										ksize=[1, sequence_length - filter_size + 1, 1, 1],
										strides=[1, 1, 1, 1],
										padding="VALID",
										name="pool")
				pooled_out.append(pooled)

		num_total_filters = num_filters * len(filter_sizes)
		self.h_pool = tf.concat(3, pooled_out)
		self.h_pool_flat = tf.reshape(self.h_pool, [-1, num_total_filters])

		# Dropout with probability: self.dropout_keep_probability
		with tf.name_scope("dropout"):
			self.h_drop = tf.nn.dropout(self.h_pool_flat, self.dropout_keep_probability)

		# Final un-normalized scores and predictions
		with tf.name_scope("output"):
			W = tf.get_variable("W", shape=[num_total_filters, num_classes],
								initializer=tf.contrib.layers.xavier_initializer())
			b = tf.Variable(tf.constant(0.1, shape=[num_classes]), name="b")
			l2_loss += tf.nn.l2_loss(W)
			l2_loss += tf.nn.l2_loss(b)
			self.scores = tf.nn.xw_plus_b(self.h_drop, W, b, name="scores")
			self.predictions = tf.argmax(self.scores, 1, name="predictions")

		# Calculate mean cross-entropy loss
		with tf.name_scope("loss"):
			# should the below be tf.nn.softmax_cross_entropy_with_logits(self.predictions, self.input_y) ???
			losses = tf.nn.softmax_cross_entropy_with_logits(self.scores, self.input_y)
			self.loss = tf.reduce_mean(losses) + l2_reg_lambda * l2_loss

		# Accuracy
		with tf.name_scope("accuracy"):
			correct = tf.equal(self.predictions, tf.argmax(self.input_y, 1))
			self.accuracy = tf.reduce_mean(tf.cast(correct, "float"), name="accuracy")