Computes ROUGE-N of two text collections of sentences. Sourece: http://research.microsoft.com/en-us/um/people/cyl/download/ papers/rouge-workin

Given a list of vocabulary file paths, return a path to a single vocabulary file that incorporates all of the vocabularies. If the list i

Adds a levelto the passed graph

Subclasses should implement this method. See the `model_fn` documentation in tf.contrib.learn.Estimator class for a more detailed explanation.

Creates the training operation

Returns the path to the configuration to be tested

Implements the logic for this bridge. This function should be implemented by child classes. Returns: A tuple initial_decoder_state ten

Creates the bridge class to be tests. Must be implemented by child classes

Creates the bridge to be used between encoder and decoder

Creates a decoder instance based on the passed parameters.

Creates and returns a new figure that visualizes attention scores for for a single model predictions.

Creates the attention layer. Should be implemented by child classes

Creates the optimizer

Create a zero outputs Tensor structure.

Runs decoding in inference mode

Runs decoding in training mode

Wraps a decoder into a Beam Search decoder. Args: decoder: The original decoder Returns: A BeamSearchDecoder with the same inter

Calcualtes n-grams. Args: n: which n-grams to calculate text: An array of tokens Returns: A set of n-grams

Returns the attention scores, sliced by source and target length.

Reads a file that specifies a mapping from source to target tokens. The file must contain lines of the form <source>\t<target>" Args: filenam

Determines whether the training has stopped.

Returns the length of the Longest Common Subsequence between sequences x and y. Source: http://www.algorithmist.com/index.php/Longest_Common_Su

Loads model from a configuration file

Parses parameter values to the types defined by the default parameters. Default parameters are used for missing values.

Model-specific preprocessing for features and labels: - Creates vocabulary lookup tables for source and target vocab - Converts tokens into v

Model-specific preprocessing for features and labels: - Creates vocabulary lookup tables for source and target vocab - Converts tokens

Logs parameter values

private recon calculation

Returns the Longest Subsequence between x and y. Source: http://www.algorithmist.com/index.php/Longest_Common_Subsequence Args: x: sequenc

Registers custom ops in the default graph. This is needed Because our checkpoint is saved with ops that are not part of Tensorflow.

Tests the encoder with a given parameter configuration

Returns the size of a tensor without the first (batch) dimension

Returns LCS_u(r_i, C) which is the LCS score of the union longest common subsequence between reference sentence ri and candidate summary C. For e

Calculates a batch- and timweise dot product with a variable

Calculates a batch- and timweise dot product