Canonical test that checks any node/variable(s) based on an evaluation model coupled with a training model. Runs training for as many it

Assigns variables passed weights Args: weight_dict: dict of variable names (under given scope) and their values

Efficiently merges arbitrary number of dicts, giving precedence to latter dicts.

Reads a tab separated value file.

Creates a `truncated_normal_initializer` with the given range.

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

Accepts a dict of dicts, each of which contains a canonical config for an initializer.

Helper function that converts canonical dict of TF ops to an actual dict. Runs ops first.

Accepts a dict in canonical config form and returns the appropriate initializer.

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

Returns the effective number of steps, i.e. number of residues that are non-missing and are not just padding, given a masking matrix. Ar

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

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 internal representation resultant from RNN output activations into dssp classes (logits).

Converts internal representation resultant from RNN or transformer output activations into geometrical representations based on one of many m

create recurrent cell(s) used in RNN

create recurrent cell(s) used in RNN

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

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

Returns a float_list from a float / double.

Gets final hidden layer of encoder. Returns: float Tensor of shape [batch_size, seq_length, hidden_size] corresponding to the fin

Creates an `input_fn` closure to be passed to TPUEstimator. Args: input_dict: Dictionary containing everything needed to feed finetun

Run layer normalization on the last dimension of the tensor.

Convert string of letters to list of ints

Indicate whether or not to enter a case suite

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

Creates alphabet for alphabetized dihedral prediction.

Recurrent layer for transforming inputs (primary sequences) into an internal representation.

Recurrent layer for transforming inputs (primary sequences) into an internal representation.

Restricts real-valued tensors to the interval [-pi, pi] by feeding them through a cosine.

See base class.

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

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

Converts pair of dicts to pair of matched tuples so that their elements can be compared. Throws an exception if the set of keys in both dic

Args: tri: NUM_DIHEDRALS x [NUM_FRAGS/0, BATCH_SIZE, NUM_DIMENSIONS] pt: [NUM_FR

Returns handles to ops (values) of nodes by name that can be evaluated by sess.run() Args: nodes: list of strings containing no

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

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

Returns an int64_list from a bool / enum / int / uint.

Computes the pairwise distance (l2 norm) between all vectors in the tensor. Vectors are assumed to be in the third dimension. Op is done ele

Computes the (possibly weighted) L2 norm of a tensor along the dimensions given in reduction_indices. Args: input_tensor: [..., NUM_DIME

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

Constructs a LAMBOptimizer.

Constructs ops to accumulate and reduce loss and maintain a memory of lowest loss achieved

Non-recurrent attention layer that examines all states, for each state, and return a convex mixture of the states.

Non-recurrent attention layer that examines all states, for each state, and return a convex mixture of the states.

Converts dihedrals into full 3D structures.

Creates TensorFlow computation graph Creates a different model depending on whether mode is set to 'training' or 'evaluation'.

Creates TF ops for maintaining and advancing the curriculum.

Creates TF queues and nodes for inputting and batching data.

Decodes a record to a TensorFlow example.

Decodes a record to a TensorFlow example.

Decodes a record to a TensorFlow example.

Returns a device placement function to insure that each operation is placed on the most optimal device.

Whether to use L2 weight decay for `param_name`.

Whether to use L2 weight decay for `param_name`.

Computes reduced weighted dRMSD loss (as specified by weights) between predicted tertiary structures and targets.

Get the variable name from the tensor name.

Get the variable name from the tensor name.

Higher-order recurrence that creates multiple layers, possibly with interleaving dihedrals and dssps

Creates op for loss history updating.

Returns final concatenated input for use in recurrent layer.

Reduces loss according to normalization order.

Computes (non-differentiable) accuracy of predicted secondary structures based on targets.

Computes reduced cross entropy loss between predicted secondary structures and targets.

Creates loss optimizer and returns minimization op.

Truncates a sequence pair in place to the maximum length.

Returns dRMSD weights that mask meaningless (missing or longer than sequence residues) pairwise distances and incorporate the state of

Raises an exception if the tensor rank is not of the expected rank. Args: tensor: A tf.Tensor to check the rank of. expected_rank: Py

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

Returns a bytes_list from a string / byte.

Create 3D attention mask from a 2D tensor mask. Args: from_tensor: 2D or 3D Tensor of shape [batch_size, from_seq_length, ...]. to_ma

Create Masked LM Instance Used to be create_instance_from_tokenized_seq

Creates several tensor outputs that are useful for fine tuning and/or prediction (inference) Returns: loss - Mean squared error loss tens

Creates a classification model.