:param inputs: [N*B, I + B] :param state: [N*B, d] :param scope: :return: [N*B, d]

:param inputs: [N, d + JQ + JQ * d] :param state: [N, d] :param scope: :return:

Early fusion attention cell: uses the (inputs, state) to control the current attention. :param cell: :param memory: [N, M, m

log_path: The full path for the log file to write. The file will be appended to if it exists. min_print_level: Only messages with lev

Calculate the average gradient for each shared variable across all towers. Note that this function provides a synchronization point across all to

A Bi-Directional LSTM layer. Returns forward and backward hidden states as a tuple, and cell states as a tuple. Ouput of hidden states: [(ba

:param inputs: [N, i] :param state: [N, d] :param sel_mem: [N, m] :return: (new_inputs, new_state) t

:param inputs: [N, i] :param state: [N, d] :param memory: [N, M, m] :return: [N, M]

Function to get accuracy and cost of the model by genre, evaluated on a chosen dataset. It returns a dictionary of accuracies by genre and cost f

To get the last hidden layer form a dynamically unrolled RNN. Input of shape (batch_size, max_seq_length, hidden_dim). true_length: Tens

Get true length of sequences (without padding), and mask for true-length in max-length. Input of shape: (batch_size, max_seq_length, hidden_

Load GloVe embeddings. Initializng OOV words to vector of zeros.

Load a specific genre's examples from MultiNLI, or load SNLI data and assign a "snli" genre to the examples. If the "snli" parameter is set t

Used to calculcate a softmax score with true sequence length (without padding), rather than max-sequence length. Input shape: (batch_size, m

Assume that inputs and sel_mem are the same size :param inputs: [N, i] :param state: [N, d] :param se

Assumes that tree span has been set with set_span Returns true if any subtree of t has exact span as the given span :param t: :param

Helper to create an initialized Variable with weight decay. Note that the Variable is initialized with a truncated normal distribution. A wei