def load_instances(instance_strs, word_vectors):
'''Load training examples
Args:
instance_strs: each string is a training example
word_vectors: an instance of vec.wordvector
Return:
instances: a list of Instance
'''
instances = [Instance.parse_from_str(i, word_vectors) for i in instance_strs]
total_internal_node = 0
for instance in instances:
total_internal_node += (len(instance.words)-1) * instance.freq
return instances, total_internal_node
def load_instances(instance_strs, word_vectors):
'''Load training examples
Args:
instance_strs: each string is a training example
word_vectors: an instance of vec.wordvector
Return:
instances: a list of Instance
'''
instances = [Instance.parse_from_str(i, word_vectors) for i in instance_strs]
total_internal_node = 0
for instance in instances:
total_internal_node += (len(instance.words)-1) * instance.freq
return instances, total_internal_node
def load_instances(instance_strs, word_vectors):
"""Load training examples
Args:
instance_strs: each string is a training example
word_vectors: an instance of vec.wordvector
Return:
instances: a list of Instance
"""
instances = [Instance.parse_from_str(i, word_vectors) for i in instance_strs]
total_internal_node = 0
for instance in instances:
# 对于一个短语有n个单词，则经过n-1次组合后形成唯一的短语向量，故中间节点共有n-1个
total_internal_node += (len(instance.words) - 1) * instance.freq
return instances, total_internal_node
print >> stderr, 'load RAE parameters...'
theta = unpickle(theta_file)
rae = RecursiveAutoencoder.build(theta, embsize)
total_cost = 0
total_instance_num = 0
total_internal_node_num = 0
print '='*63
print '%20s %20s %20s' % ('all', 'avg/node', 'internal node')
print '-'*63
with Reader(phrases_file) as reader, Writer(output_file) as writer:
for phrase in reader:
instance = Instance.parse_from_str(phrase, word_vectors)
words_embedded = word_vectors[instance.words]
root_node, cost = rae.forward(words_embedded)
# print "root node: ",root_node
# vec = root_node.p.T[0] # convert n*1 vector to common vector
vec=[]
for node in root_node:
vec.append(node.p.T[0])
# for nodes in root_node:
# print "n:",nodes.p.shape
# continue
# writer.write(' '.join([str(vec[i]) for i in range(vec.size)]))
for j in range(len(vec)):
v=vec[j]
writer.write(' '.join([str(v[i]) for i in range(v.size)]))
if not j==(len(vec)-1):
print >> stderr, 'load RAE parameters...'
theta = unpickle(theta_file)
rae = RecursiveAutoencoder.build(theta, embsize)
total_cost = 0
total_instance_num = 0
total_internal_node_num = 0
print '=' * 63
print '%20s %20s %20s' % ('all', 'avg/node', 'internal node')
print '-' * 63
with Reader(phrases_file) as reader, Writer(output_file) as writer:
for phrase in reader:
instance = Instance.parse_from_str(phrase, word_vectors)
words_embedded = word_vectors[instance.words]
root_node, cost = rae.forward(words_embedded)
vec = root_node.p.T[0] # convert n*1 vector to common vector
writer.write(' '.join([str(vec[i]) for i in range(vec.size)]))
writer.write('\n')
internal_node_num = len(instance.words) - 1
if internal_node_num > 0:
print '%20.8f, %20.8f, %18d' % (cost, cost / internal_node_num,
internal_node_num)
else:
print '%20.8f, %20.8f, %18d' % (cost, cost, 0)
total_cost += cost
total_instance_num += 1
