Run decoding mode.

Generator for the training and evaluation data. Generate source and target data from a single file. Args: data_path: The path to data f

Prepare utilities for decoding.

Run tensorflow operation for decoding.

Create dictionary mapping word to its different pronounciations. Args: source_path: path to the data file; cleanup: flag indicating whether

Compute predictions on entries in filename and write them out.

Initialize vocabulary with sym from sym_generator.

Calculate a number of prediction errors.

Run evaluation mode.

Generate cases from generator and save as TFRecord file. Args: generator: a generator yielding (string -> int/float/str list) data. output_

Get next word in the interactive mode.

Split line into graphemes and phonemes. Args: input_line: raw input line; cleanup: flag indicating whether to cleanup datasets from stress

Run training.

Initialize generator.

Load freezed graph.

Create monitored session and generator that reads from the terminal and yields "interactive inputs". Due to temporary limitations in tf.learn

Decode batch

Returning inputs. Args: filename: path to file with inputs, 1 per line. delimiters: str, delimits records in the file. Returns: a li

Load vocab from a file. Args: filename: The file to load vocabulary from.

Initialize symbols from a list of symbols. It is ok if reserved symbols appear in the vocab list. They will be removed. The set of symbols in

Add problem hparams for the problems.

Reads a file to build a vocabulary with letters and phonemes. Args: data_path: data file to read list of words from. Returns: vo

Create train, development and test data files from initial data files in case when not provided development or test data files or active cleanup f

Create Experiment.

Wrapper for canonical experiment_fn. See create_experiment.

Create grapheme-to-phoneme ground true mapping.

Decode word. Args: word: word for decoding. Returns: pronunciation: a decoded phonemes sequence for input word.

Freeze pre-trained model.

Generate and save preprocessed data as TFRecord files. Args: train_path: the path to the train data file. eval_path: the path to the

Generate cases from a generators and save as TFRecord files. Generated cases are transformed to tf.Example protos and saved as TFRecords in shard

Interactive decoding.

Write vocab file to disk. Vocab files have one symbol per line. The file ends in a newline. Reserved symbols are written to the vocab file as

r"""Generator for sequence-to-sequence tasks using tabbed files. Tokens are derived from text files where each line contains both a source an

Create a Problem. Args: was_reversed: bool, whether to reverse inputs and targets. was_copy: bool, whether to copy inputs to targets.

Initialize from a file or list, one token per line. Handling of reserved tokens works as follows: - When initializing from a list, we add res

Whether preprocess data into required format.

Create RunConfig

Build a Dataset for this problem. Args: mode: tf.estimator.ModeKeys; determines which files to read from. data_dir: directory that co

Serialized Example to dict of <feature name, Tensor>.

Input function returning features which is a dictionary of string feature name to `Tensor` or `SparseTensor`. If it returns a tuple, f

Load/create vocabularies.

Load customizable parameters from 'model.params' file.

Main function.

Save customizable model parameters in 'model.params' file.

Symbols generator for vocab initializer from file.