add my_seq2seq_v2.py

lichuang · lichuang · commit a18d332b3b3f · 2017-01-09T08:06:48.000+08:00
diff --git a/chatbotv2/my_seq2seq_v2.py b/chatbotv2/my_seq2seq_v2.py
@@ -0,0 +1,250 @@
+# -*- coding: utf-8 -*-
+
+import sys
+import math
+import tflearn
+import tensorflow as tf
+from tensorflow.python.ops import rnn_cell
+from tensorflow.python.ops import rnn
+import chardet
+import numpy as np
+import struct
+
+question_seqs = []
+answer_seqs = []
+
+max_w = 50
+float_size = 4
+word_vector_dict = {}
+trained_word_vector_dict = {}
+word_vec_dim = 200
+max_seq_len = 8
+word_set = {}
+
+def load_word_set():
+    file_object = open('./segment_result_lined.3000000.pair.less', 'r')
+    while True:
+        line = file_object.readline()
+        if line:
+            line_pair = line.split('|')
+            line_question = line_pair[0]
+            line_answer = line_pair[1]
+            for word in line_question.decode('utf-8').split(' '):
+                word_set[word] = 1
+            for word in line_answer.decode('utf-8').split(' '):
+                word_set[word] = 1
+        else:
+            break
+    file_object.close()
+
+def load_vectors(input):
+    """从vectors.bin加载词向量，返回一个word_vector_dict的词典，key是词，value是200维的向量
+    """
+    print "begin load vectors"
+
+    input_file = open(input, "rb")
+
+    # 获取词表数目及向量维度
+    words_and_size = input_file.readline()
+    words_and_size = words_and_size.strip()
+    words = long(words_and_size.split(' ')[0])
+    size = long(words_and_size.split(' ')[1])
+    print "words =", words
+    print "size =", size
+
+    for b in range(0, words):
+        a = 0
+        word = ''
+        # 读取一个词
+        while True:
+            c = input_file.read(1)
+            word = word + c
+            if False == c or c == ' ':
+                break
+            if a < max_w and c != '\n':
+                a = a + 1
+        word = word.strip()
+
+        vector = []
+        for index in range(0, size):
+            m = input_file.read(float_size)
+            (weight,) = struct.unpack('f', m)
+            vector.append(float(weight))
+
+        # 将词及其对应的向量存到dict中
+
+        if word_set.has_key(word.decode('utf-8')):
+            word_vector_dict[word.decode('utf-8')] = vector[0:word_vec_dim]
+
+    input_file.close()
+
+    print "load vectors finish"
+
+def init_seq(input_file):
+    """读取切好词的文本文件，加载全部词序列
+    """
+    file_object = open(input_file, 'r')
+    vocab_dict = {}
+    while True:
+        question_seq = []
+        answer_seq = []
+        line = file_object.readline()
+        if line:
+            line_pair = line.split('|')
+            line_question = line_pair[0]
+            line_answer = line_pair[1]
+            for word in line_question.decode('utf-8').split(' '):
+                if word_vector_dict.has_key(word):
+                    question_seq.append(word_vector_dict[word])
+                    trained_word_vector_dict[word] = word_vector_dict[word]
+            for word in line_answer.decode('utf-8').split(' '):
+                if word_vector_dict.has_key(word):
+                    answer_seq.append(word_vector_dict[word])
+                    trained_word_vector_dict[word] = word_vector_dict[word]
+        else:
+            break
+        question_seqs.append(question_seq)
+        answer_seqs.append(answer_seq)
+    file_object.close()
+
+def vector_sqrtlen(vector):
+    len = 0
+    for item in vector:
+        len += item * item
+    len = math.sqrt(len)
+    return len
+
+def vector_cosine(v1, v2):
+    if len(v1) != len(v2):
+        sys.exit(1)
+    sqrtlen1 = vector_sqrtlen(v1)
+    sqrtlen2 = vector_sqrtlen(v2)
+    value = 0
+    for item1, item2 in zip(v1, v2):
+        value += item1 * item2
+    return value / (sqrtlen1*sqrtlen2)
+
+
+def vector2word(vector):
+    max_cos = -10000
+    match_word = ''
+    for word in word_vector_dict:
+        v = word_vector_dict[word]
+        cosine = vector_cosine(vector, v)
+        if cosine > max_cos:
+            max_cos = cosine
+            match_word = word
+    return (match_word, max_cos)
+
+
+class MySeq2Seq(object):
+    """
+    思路：输入输出序列一起作为input，然后通过slick和unpack切分
+    完全按照论文说的编码器解码器来做
+    输出的时候把解码器的输出按照词向量的200维展平，这样输出就是(?,seqlen*200)
+    这样就可以通过regression来做回归计算了，输入的y也展平，保持一致
+    """
+    def __init__(self, max_seq_len = 16, word_vec_dim = 200, input_file='./segment_result_lined.3000000.pair.less'):
+        self.max_seq_len = max_seq_len
+        self.word_vec_dim = word_vec_dim
+        self.input_file = input_file
+
+    def generate_trainig_data(self):
+        load_word_set()
+        load_vectors("./vectors.bin")
+        init_seq(self.input_file)
+        xy_data = []
+        y_data = []
+        for i in range(len(question_seqs)):
+        #for i in range(100):
+            question_seq = question_seqs[i]
+            answer_seq = answer_seqs[i]
+            if len(question_seq) < self.max_seq_len and len(answer_seq) < self.max_seq_len:
+                sequence_xy = [np.zeros(self.word_vec_dim)] * (self.max_seq_len-len(question_seq)) + list(reversed(question_seq))
+                sequence_y = answer_seq + [np.zeros(self.word_vec_dim)] * (self.max_seq_len-len(answer_seq))
+                sequence_xy = sequence_xy + sequence_y
+                sequence_y = [np.ones(self.word_vec_dim)] + sequence_y
+                xy_data.append(sequence_xy)
+                y_data.append(sequence_y)
+
+                #print "right answer"
+                #for w in answer_seq:
+                #    (match_word, max_cos) = vector2word(w)
+                #    if len(match_word)>0:
+                #        print match_word, vector_sqrtlen(w)
+
+        return np.array(xy_data), np.array(y_data)
+
+
+    def model(self, feed_previous=False):
+        # 通过输入的XY生成encoder_inputs和带GO头的decoder_inputs
+        input_data = tflearn.input_data(shape=[None, self.max_seq_len*2, self.word_vec_dim], dtype=tf.float32, name = "XY")
+        encoder_inputs = tf.slice(input_data, [0, 0, 0], [-1, self.max_seq_len, self.word_vec_dim], name="enc_in")
+        decoder_inputs_tmp = tf.slice(input_data, [0, self.max_seq_len, 0], [-1, self.max_seq_len-1, self.word_vec_dim], name="dec_in_tmp")
+        go_inputs = tf.ones_like(decoder_inputs_tmp)
+        go_inputs = tf.slice(go_inputs, [0, 0, 0], [-1, 1, self.word_vec_dim])
+        decoder_inputs = tf.concat(1, [go_inputs, decoder_inputs_tmp], name="dec_in")
+
+        # 编码器
+        # 把encoder_inputs交给编码器，返回一个输出(预测序列的第一个值)和一个状态(传给解码器)
+        (encoder_output_tensor, states) = tflearn.lstm(encoder_inputs, self.word_vec_dim, return_state=True, scope='encoder_lstm')
+        encoder_output_sequence = tf.pack([encoder_output_tensor], axis=1)
+
+        # 解码器
+        # 预测过程用前一个时间序的输出作为下一个时间序的输入
+        # 先用编码器的最后一个输出作为第一个输入
+        if feed_previous:
+            first_dec_input = go_inputs
+        else:
+            first_dec_input = tf.slice(decoder_inputs, [0, 0, 0], [-1, 1, self.word_vec_dim])
+        decoder_output_tensor = tflearn.lstm(first_dec_input, self.word_vec_dim, initial_state=states, return_seq=False, reuse=False, scope='decoder_lstm')
+        decoder_output_sequence_single = tf.pack([decoder_output_tensor], axis=1)
+        decoder_output_sequence_list = [decoder_output_tensor]
+        # 再用解码器的输出作为下一个时序的输入
+        for i in range(self.max_seq_len-1):
+            if feed_previous:
+                next_dec_input = decoder_output_sequence_single
+            else:
+                next_dec_input = tf.slice(decoder_inputs, [0, i+1, 0], [-1, 1, self.word_vec_dim])
+            decoder_output_tensor = tflearn.lstm(next_dec_input, self.word_vec_dim, return_seq=False, reuse=True, scope='decoder_lstm')
+            decoder_output_sequence_single = tf.pack([decoder_output_tensor], axis=1)
+            decoder_output_sequence_list.append(decoder_output_tensor)
+
+        decoder_output_sequence = tf.pack(decoder_output_sequence_list, axis=1)
+        real_output_sequence = tf.concat(1, [encoder_output_sequence, decoder_output_sequence])
+
+        net = tflearn.regression(real_output_sequence, optimizer='sgd', learning_rate=0.1, loss='mean_square')
+        model = tflearn.DNN(net)
+        return model
+
+    def train(self):
+        trainXY, trainY = self.generate_trainig_data()
+        model = self.model(feed_previous=False)
+        model.fit(trainXY, trainY, n_epoch=1000, snapshot_epoch=False, batch_size=1)
+        model.save('./model/model')
+        return model
+
+    def load(self):
+        model = self.model(feed_previous=True)
+        model.load('./model/model')
+        return model
+
+if __name__ == '__main__':
+    phrase = sys.argv[1]
+    if 3 == len(sys.argv):
+        my_seq2seq = MySeq2Seq(word_vec_dim=word_vec_dim, max_seq_len=max_seq_len, input_file=sys.argv[2])
+    else:
+        my_seq2seq = MySeq2Seq(word_vec_dim=word_vec_dim, max_seq_len=max_seq_len)
+    if phrase == 'train':
+        my_seq2seq.train()
+    else:
+        model = my_seq2seq.load()
+        trainXY, trainY = my_seq2seq.generate_trainig_data()
+        predict = model.predict(trainXY)
+        for sample in predict:
+            print "predict answer"
+            for w in sample[1:]:
+                (match_word, max_cos) = vector2word(w)
+                #if vector_sqrtlen(w) < 1:
+                #    break
+                print match_word, max_cos, vector_sqrtlen(w)
diff --git a/chatbotv2/readme.txt b/chatbotv2/readme.txt
@@ -0,0 +1,9 @@
+python ../word_segment.py zhenhuanzhuan.txt zhenhuanzhuan.segment
+../word2vec/word2vec -train ./zhenhuanzhuan.segment -output vectors.bin -cbow 1 -size 200 -window 8 -negative 25 -hs 0 -sample 1e-4 -threads 20 -binary 1 -iter 15
+
+
+
+head -10000 ../subtitle/raw_subtitles/subtitle.corpus > subtitle.corpus.10000
+python ../word_segment.py subtitle.corpus.10000 subtitle.corpus.10000.segment
+../word2vec/word2vec -train ./subtitle.corpus.10000.segment -output vectors.bin -cbow 1 -size 200 -window 8 -negative 25 -hs 0 -sample 1e-5 -threads 20 -binary 1 -iter 15
+cat subtitle.corpus.10000.segment | awk '{if(last!="")print last"|"$0;last=$0}' | sed 's/| /|/g' > subtitle.corpus.10000.segment.pair
