See base class.

See base class.

See base class.

See base class.

See base class.

See base class.

See base class.

Gets final hidden layer of encoder. Returns: float Tensor of shape [batch_size, seq_length, hidden_size] corresponding to the final h

Gets a collection of `InputExample`s for prediction.

See base class.

See base class.

See base class.

See base class.

See base class.

Gets a collection of `InputExample`s for the train set.

See base class.

See base class.

See base class.

See base class.

See base class.

See base class.

get weights for current batch :param answer_list: a numpy array contain labels for a batch :param weights_dict: a dict that contain

single step of gru for word level :param Xt: Xt:[batch_size,embed_size] :param h_t_minus_1:[batch_size,embed_size] :r

single step of gru for word level :param Xt: Xt:[batch_size,embed_size] :param h_t_minus_1:[batch_size,embed_size] :p

The actual input function.

Creates an `input_fn` closure to be passed to TPUEstimator.

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

load data from h5py and pickle cache files, which is generate by take step by step of pre-processing.ipynb :param cache_file_h5py: :param

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list repre

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

convert data as indexes using word2index dicts. :param traning_data_path: :param vocab_word2index: :param vocab_label2index: :ret

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list repre

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list repre

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list repre

input: a file path :return: train, test, valid. where train=(trainX, trainY). where trainX: is a list of list.each list re

calculate loss using (NCE)cross entropy here

calculate loss using (NCE)cross entropy here

The `model_fn` for TPUEstimator.

Returns text encoded in a way suitable for print or `tf.logging`.

:param row: it is a list.length is number of labels. e.g. 2002 :param vocabulary_index2word_label :param result_list :return: a l

RNN decoder for the sequence-to-sequence model. Args: decoder_inputs: A list of 2D Tensors [batch_size x input_size].it is target Y, but s

scaled dot product attention :param Q: query. shape:[batch,sequence_length,d_model] :param K_s: keys. shape:[batch,sequ

Cross entropy with label smoothing to limit over-confidence.