def __init__(self):
tf.reset_default_graph()
self.encoder_vec_file = "./tfdata/enc.vec"
self.decoder_vec_file = "./dec.vec"
self.encoder_vocabulary = "./tfdata/enc.vocab"
self.decoder_vocabulary = "./tfdata/dec.vocab"
self.batch_size = 1
self.max_batches = 100000
self.show_epoch = 100
self.model_path = './model/'
self.model = dynamicSeq2seq(encoder_cell=LSTMCell(40),
decoder_cell=LSTMCell(40),
encoder_vocab_size=600,
decoder_vocab_size=1600,
embedding_size=20,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
self.location = ["杭州", "重庆", "上海", "北京"]
self.dec_vocab = {}
self.enc_vocab = {}
self.dec_vecToSeg = {}
tag_location = ''
with open(self.encoder_vocabulary, "r") as enc_vocab_file:
for index, word in enumerate(enc_vocab_file.readlines()):
self.enc_vocab[word.strip()] = index
with open(self.decoder_vocabulary, "r") as dec_vocab_file:
for index, word in enumerate(dec_vocab_file.readlines()):
self.dec_vecToSeg[index] = word.strip()
self.dec_vocab[word.strip()] = index
def __init__(self):
print("tensorflow version: ", tf.__version__)
tf.reset_default_graph()
self.encoder_vec_file = "./preprocessing/encode.vector"
self.decoder_vec_file = "./preprocessing/decode.vector"
self.encoder_vocabulary = "./preprocessing/encode.vocabulary"
self.decoder_vocabulary = "./preprocessing/decode.vocabulary"
self.batch_size = 1
self.max_batches = 10000
self.show_epoch = 1000
self.model_path = './model/'
self.model = dynamicSeq2seq(encoder_cell=LSTMCell(40),
decoder_cell=LSTMCell(40),
encoder_vocab_size=600,
decoder_vocab_size=1600,
embedding_size=20,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
self.dec_vocab = {}
self.enc_vocab = {}
self.dec_vecToSeg = {}
tag_location = ''
with io.open(self.encoder_vocabulary, "r",
encoding="utf-8") as enc_vocab_file:
for index, word in enumerate(enc_vocab_file.readlines()):
self.enc_vocab[word.strip()] = index
with io.open(self.decoder_vocabulary, "r",
encoding="utf-8") as dec_vocab_file:
for index, word in enumerate(dec_vocab_file.readlines()):
self.dec_vecToSeg[index] = word.strip()
self.dec_vocab[word.strip()] = index
def __init__(self):
tf.reset_default_graph()
self.encoder_vec_file = "./tfdata/enc.vec"
self.decoder_vec_file = "./tfdata/dec.vec"
self.encoder_vocabulary = "./tfdata/enc.vocab"
self.decoder_vocabulary = "./tfdata/dec.vocab"
self.batch_size = 1
self.max_batches = 100000
self.show_epoch = 100
self.model_path = './model/'
self.model = dynamicSeq2seq(encoder_cell=LSTMCell(40),
decoder_cell=LSTMCell(40),
encoder_vocab_size=600,
decoder_vocab_size=1600,
embedding_size=20,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
self.location = ["杭州", "重庆", "上海", "北京"]
self.dec_vocab = {}
self.enc_vocab = {}
self.dec_vecToSeg = {}
tag_location = ''
with open(self.encoder_vocabulary, "r") as enc_vocab_file:
for index, word in enumerate(enc_vocab_file.readlines()):
self.enc_vocab[word.strip()] = index
with open(self.decoder_vocabulary, "r") as dec_vocab_file:
for index, word in enumerate(dec_vocab_file.readlines()):
self.dec_vecToSeg[index] = word.strip()
self.dec_vocab[word.strip()] = index
def __init__(self):
tf.reset_default_graph()
self.encoder_sege_file = "./tf_data_new/enc.segement"
self.decoder_sege_file = "./tf_data_new/dec.segement"
self.encoder_vocabulary = "./tf_data_new/enc.vocab"
self.decoder_vocabulary = "./tf_data_new/dec.vocab"
self.eval_enc = "./tf_data_new/eval_enc"
self.eval_dec = "./tf_data_new/eval_dec"
self.vocab_file = "./tf_data_new/en_de_vocabs"
self.batch_size = 20
self.max_batches = 15000
self.show_epoch = 10
self.model_path = './model_2/'
self.transform_model = Transformer(
embedding_size=128,
num_layers=6,
keep_prob_rate=0.2,
learning_rate=0.0001,
learning_decay_rate=0.99,
clip_gradient=True,
is_embedding_scale=True,
multihead_num=8,
max_gradient_norm=5,
vocab_size=40020,
max_encoder_len=200,
max_decoder_len=200,
share_embedding=True,
pad_index=0,
learning_decay_steps=500,
dimension_feedforword=2048,
dimension_model=512,
)
self.LSTMmodel = dynamicSeq2seq(encoder_cell=LSTMCell(500),
decoder_cell=LSTMCell(500),
encoder_vocab_size=70824,
decoder_vocab_size=70833,
embedding_size=128,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
def __init__(self):
print("tensorflow version: ", tf.__version__)
tf.reset_default_graph()
self.encoder_vec_file = './preprocessing/enc.vec'
self.decoder_vec_file = './preprocessing/dec.vec'
self.encoder_vocabulary = './preprocessing/enc.vocab'
self.decoder_vocabulary = './preprocessing/dec.vocab'
self.dictFile = './word_dict.txt'
self.batch_size = 1
self.max_batches = 100000
self.show_epoch = 100
self.model_path = './model/'
# jieba导入词典
jieba.load_userdict(self.dictFile)
self.model = dynamicSeq2seq(encoder_cell=LSTMCell(40),
decoder_cell=LSTMCell(40),
encoder_vocab_size=600,
decoder_vocab_size=1600,
embedding_size=20,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
self.location = ["杭州", "重庆", "上海", "北京"]
self.user_info = {"__username__": "yw", "__location__": "重庆"}
self.robot_info = {"__robotname__": "Rr"}
self.dec_vocab = {}
self.enc_vocab = {}
self.dec_vecToSeg = {}
tag_location = ''
with open(self.encoder_vocabulary, "r") as enc_vocab_file:
for index, word in enumerate(enc_vocab_file.readlines()):
self.enc_vocab[word.strip()] = index
with open(self.decoder_vocabulary, "r") as dec_vocab_file:
for index, word in enumerate(dec_vocab_file.readlines()):
self.dec_vecToSeg[index] = word.strip()
self.dec_vocab[word.strip()] = index
def __init__(self):
tf.reset_default_graph() # 用于清除默认图形堆栈并重置全局默认图形。
self.encoder_vec_file = "./tfdata/enc.vec"
self.decoder_vec_file = "./tfdata/dec.vec"
self.encoder_vocabulary = "./tfdata/enc.vocab"
self.decoder_vocabulary = "./tfdata/dec.vocab"
self.batch_size = 1
self.max_batches = 100000
self.show_epoch = 100
self.model_path = './model/'
self.model = dynamicSeq2seq(encoder_cell=LSTMCell(40),
decoder_cell=LSTMCell(40),
encoder_vocab_size=600,
decoder_vocab_size=1600,
embedding_size=20,
attention=False,
bidirectional=False,
debug=False,
time_major=True)
self.location = ["杭州", "重庆", "上海", "北京"]
self.dec_vocab = {}
self.enc_vocab = {}
self.dec_vecToSeg = {}
tag_location = ''
with open(self.encoder_vocabulary, "r") as enc_vocab_file:
# enumerate用于将一个可遍历的数据对象(如列表、元组或字符串)组合为一个索引序列，同时列出数据和数据下标
# readlines用于读取所有行(直到结束符EOF)并返回列表
for index, word in enumerate(enc_vocab_file.readlines()):
# strip用于移除字符串头尾指定的字符（默认为空格或换行符）或字符序列
self.enc_vocab[word.strip()] = index
with open(self.decoder_vocabulary, "r") as dec_vocab_file:
for index, word in enumerate(dec_vocab_file.readlines()):
self.dec_vecToSeg[index] = word.strip()
self.dec_vocab[word.strip()] = index