Reads a tab separated value file.

Creates a `truncated_normal_initializer` with the given range.

Perform dropout. Args: input_tensor: float Tensor. dropout_prob: Python float. The probability of dropping out a value (NOT of *keepi

Returns a list of the shape of tensor, preferring static dimensions. Args: tensor: A tf.Tensor object to find the shape of. expected_rank:

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

layer norm & residual connection :param input: [batch_size,equence_length,d_model] :param output:[batch_size,sequence_leng

Converts a sequence of [tokens|ids] using the vocab.

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

multi head attention as sub layer :param layer_index: index of layer number :param Q: shape should be: [batch_size,sequenc

Creates examples for the training and dev sets.

Creates examples for the training and dev sets.

Creates examples for the training and dev sets.

Creates examples for the training and dev sets.

Creates examples for the training and dev sets.

Converts `text` to Unicode (if it's not already), assuming utf-8 input.

Convert a set of `InputExample`s to a TFRecord file.

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

Run layer normalization on the last dimension of the tensor.

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

multi head attention :param Q: query. shape:[batch,sequence_length,d_model] :param K_s: keys. shape:[batch,sequence_lengt

Reshapes a >= rank 2 tensor to a rank 2 tensor (i.e., a matrix).

:param x: shape should be:[batch_size,sequence_length,d_model] :param layer_index: index of layer number :param type: enco

based on the loss, use SGD to update parameter

Runs basic whitespace cleaning and splitting on a peice of text.

See base class.

:param input_sequence: [batch_size,seq_length,num_units] :param attention_level: word or sentence level :return: [batch_size,

:param input_sequences: [batch_size,seq_lenght,num_units] :param level: word or sentence :param reuse_flag: resuse or not

compute confuse matrix for a batch :param y_targetlabel_list: a list; each element is a mulit-hot,e.g. [1,0,0,1,...] :param y_logits_arra

compute f1 :param TP_micro: number.e.g. 200 :param FP_micro: number.e.g. 200 :param FN_micro: number.e.g. 200 :return: f1_score:

compute f1 of micro and macro :param label_dict: :return: f1_micro,f1_macro: scalar, scalar

Converts a single `InputExample` into a single `InputFeatures`.

singel layer for decoder. each layers has three sub-layers: the first is multi-head self-attention(mask) mechanism; the se

Constructs a `BertConfig` from a json file of parameters.

get input mask and segment ids given a batch of input x. if sequence length of input x is max_sequence_length, then shape of both input_mask

get input mask and segment ids given a batch of input x. if sequence length of input x is max_sequence_length, then shape of both input_mask

single step of gru :param Xt: Xt:[batch_size,hidden_size] :param h_t_minus_1:[batch_size,hidden_size] :return:[ba

:param embedded_words:[None,sequence_length, self.embed_size] :return:forward hidden state: a list.length is sentence_length, each el

single step of gru for sentence level :param Xt:[batch_size, hidden_size*2] :param h_t_minus_1:[batch_size, hidden_size*2]

single step of gru for sentence level :param Xt:[batch_size, hidden_size*2] :param h_t_minus_1:[batch_size, hidden_size*2]

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

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

init label dict. this dict will be used to save TP,FP,FN :param num_classes: :return: label_dict: a dict. {label_index:(0,0,0)}

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

x: [batch,sequence_length,d_model] :return: [batch,sequence_length,d_model]

Reshapes a rank 2 tensor back to its original rank >= 2 tensor.

:param label_list: e.g.[0,1,4] :param label_size: e.g.199 :return:e.g.[1,1,0,1,0,0,........]

:param label_list: e.g.[0,1,4] :param label_size: e.g.199 :return:e.g.[1,1,0,1,0,0,........]

:param label_list: e.g.[0,1,4] :param label_size: e.g.199 :return:e.g.[1,1,0,1,0,0,........]

:param label_list: e.g.[0,1,4] :param label_size: e.g.199 :return:e.g.[1,1,0,1,0,0,........]