String that translates size in bytes to kB, Mb, etc.

Serializes this instance to a JSON string.

Returns a list of Tokens.

Attempts to execute a SQL query against a database given a cursor.

Union two sketches. Args: sk1: first sketch sk2: second sketch Returns: a countmin sketch for union

Weight elements in some set by the sum of the scores in some other set. Args: other: A NeuralQueryExpression Returns: The Neural

Compute feature representations for attention/copy. Args: hps: hyperparameters. encoder_outputs: outputs by the encoder RNN. Returns:

Compute feature representations for attention/copy. Args: hps: hyperparameters. encoder_outputs: outputs by the encoder RNN. Returns:

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.

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.

Create test examples for fever task. This function is similar to create_validation_examples, but handles the fact that: (1) there are no gold

Create validation examples for fever task. This function follows these steps/guidelines: 1. Get up to the top n_inference_documents rated doc

Finds a reference to a table in the schema.

Generates 2D index given separate arrays for each coordinate.

Return set of non-terminal symbols in `rhs`.

Returns SqlTable tuple for the given token_value.

Like `_get_arg`, but also resolves `tuple_index`.

Get the i'th hash value of element x. Args: x: name or id in string i: the i'th hash function Returns: hash result

Returns true for dynamic dimensions (a tensor value, not an int).

Logs the statistics as percentages of the total count.

Parses lines from UNCorpus dataset.

Parses lines from IWSLT17 dataset.

Read reference file.

Helper function to append shard_id to the end of rel_name. Args: rel_name: string naming a declared relation shard_id: the i'th shard

Run a bi-directional LSTM with highway connections over `x`. Args: x: <tf.float32>[batch, seq_len, dim] seq_len: <tf.int32>[batch] for None

Performs multi-headed attention from `from_tensor` to `to_tensor`. This is an implementation of multi-headed attention based on "Attention is all

Adds base hparams for AgreementMultilingualNmt.

Adds base attention hparams.

Base hparams for LSTM-based NMT models.

String name for a cell in the grid.

Clean word pieces.

Tokens are lower-cased, stripped of punctuation, and filtered for NULL.

Compute logits for word prediction.

Return tf.Example or Raise.

Convert evidence inputs to their string representation. In particular, this means output looks like: Albert Einstein @@septitle@@ @@sid_1@@ the e

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

Padds or truncates 1D tensor to a specified length.

Like re.findall but returns ints instead of strings.

Freeze the SymbolTable for a specific type. Args: type_name: string name of a type. unknown_marker: string Unknown words or words ad

Constructs a `TransformerConfig` from a Python dict of parameters.

Init JointSampler from QCFG rules and Target CFG rules. Args: target_grammar_rules: A list of TargetCfgRule instance. qcfg_rules: A l

Get entries based on full key.

Return information associated with a declared relation group. Args: group_name: a string naming a group Returns: The RelationGro

Create the lookup table base on the vocabulary.

Computes metrics over precision and recall for a specific groups.

Gets the vocab id of `key`.

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

Test if the type is declared as a relation group. Args: group_name: a string possible naming a group Returns: boolean

Test if the relation has been declared as 'trainable'. Args: rel_name: a string naming a declared relation. Returns: a boolean

Load model weights from file. Args: config: experiment config. Returns: Dictionary of model weights.

Collect metrics for function.

Returns `model_fn` closure for TPUEstimator.

Return positions of non-zero elements in the input 1D tensor.

Unicode normalize the wikipedia title. Args: wikipedia_url: The original title Returns: The unicode normalized title