SetTransformer

class bayesflow.networks.SetTransformer(*args, **kwargs)

Bases: SummaryNetwork

Implements the set transformer architecture from [1] which ultimately represents a learnable permutation-invariant function. Designed to naturally model interactions in the input set, which may be hard to capture with the simpler DeepSet architecture.

[1] Lee, J., Lee, Y., Kim, J., Kosiorek, A., Choi, S., & Teh, Y. W. (2019).

Set transformer: A framework for attention-based permutation-invariant neural networks. In International conference on machine learning (pp. 3744-3753). PMLR.

Note: Currently works only on 3D inputs but can easily be expanded by using keras.layers.TimeDistributed.

Creates a many-to-one permutation-invariant encoder, typically used as a summary net for embedding set-based, (i.e., exchangeable or IID) data. Use a TimeSeriesTransformer or a FusionTransformer for non-IID data.

The number of multi-head attention block is inferred from the length of embed_dims tuple.

Parameters:

summary_dimint, optional (default - 16)

Dimensionality of the final summary output.

embed_dimstuple of int, optional (default - (64, 64))

Dimensions of the keys, values, and queries for each attention block.

num_headstuple of int, optional (default - (4, 4))

Number of attention heads for each embedding dimension.

mlp_depthstuple of int, optional (default - (2, 2))

Depth of the multi-layer perceptron (MLP) blocks for each component.

mlp_widthstuple of int, optional (default - (128, 128))

Width of each MLP layer in each block for each component.

num_seedsint, optional (default - 1)

Number of seeds to use for embedding.

dropoutfloat, optional (default - 0.05)

Dropout rate applied to the attention and MLP layers. If set to None, no dropout is applied.

mlp_activationstr, optional (default - ‘gelu’)

Activation function used in the dense layers. Common choices include “relu”, “elu”, and “gelu”.

kernel_initializerstr, optional (default - ‘he_normal’)

Initializer for the kernel weights matrix. Common choices include “glorot_uniform”, “he_normal”, etc.

use_biasbool, optional (default - True)

Whether to include a bias term in the dense layers.

layer_normbool, optional (default - True)

Whether to apply layer normalization after the attention and MLP layers.

num_inducing_pointsint or None, optional (default - None)

Number of inducing points used, if applicable. If set to None, this option is disabled.

seed_dimint or None, optional (default - None)

Dimensionality of the seed embeddings. If None, it defaults to summary_dim.

**kwargsdict

Additional keyword arguments passed to the base layer.

call(input_set: Tensor, training: bool = False, **kwargs) → Tensor

Compresses the input sequence into a summary vector of size summary_dim.

Parameters:

input_setTensor (e.g., np.ndarray, tf.Tensor, …)

Input of shape (batch_size, set_size, input_dim)

trainingboolean, optional (default - False)

Passed to the optional internal dropout and spectral normalization layers to distinguish between train and test time behavior.

**kwargsdict, optional (default - {})

Additional keyword arguments passed to the internal attention layer, such as attention_mask or return_attention_scores

Returns:

outTensor

Output of shape (batch_size, set_size, output_dim)

build(input_shape)

__call__(*args, **kwargs)

Call self as a function.

add_loss(loss)

Can be called inside of the call() method to add a scalar loss.

Example:

class MyLayer(Layer):
    ...
    def call(self, x):
        self.add_loss(ops.sum(x))
        return x

add_metric(*args, **kwargs)

add_variable(shape, initializer, dtype=None, trainable=True, autocast=True, regularizer=None, constraint=None, name=None)

Add a weight variable to the layer.

Alias of add_weight().

add_weight(shape=None, initializer=None, dtype=None, trainable=True, autocast=True, regularizer=None, constraint=None, aggregation='none', name=None)

Add a weight variable to the layer.

Args:

shape: Shape tuple for the variable. Must be fully-defined

(no None entries). Defaults to () (scalar) if unspecified.

initializer: Initializer object to use to populate the initial

variable value, or string name of a built-in initializer (e.g. “random_normal”). If unspecified, defaults to “glorot_uniform” for floating-point variables and to “zeros” for all other types (e.g. int, bool).

dtype: Dtype of the variable to create, e.g. “float32”. If

unspecified, defaults to the layer’s variable dtype (which itself defaults to “float32” if unspecified).

trainable: Boolean, whether the variable should be trainable via

backprop or whether its updates are managed manually. Defaults to True.

autocast: Boolean, whether to autocast layers variables when

accessing them. Defaults to True.

regularizer: Regularizer object to call to apply penalty on the

weight. These penalties are summed into the loss function during optimization. Defaults to None.

constraint: Contrainst object to call on the variable after any

optimizer update, or string name of a built-in constraint. Defaults to None.

aggregation: Optional string, one of None, “none”, “mean”,

“sum” or “only_first_replica”. Annotates the variable with the type of multi-replica aggregation to be used for this variable when writing custom data parallel training loops. Defaults to “none”.

name: String name of the variable. Useful for debugging purposes.

build_from_config(config)

Builds the layer’s states with the supplied config dict.

By default, this method calls the build(config[“input_shape”]) method, which creates weights based on the layer’s input shape in the supplied config. If your config contains other information needed to load the layer’s state, you should override this method.

Args:

config: Dict containing the input shape associated with this layer.

property compute_dtype

The dtype of the computations performed by the layer.

compute_mask(inputs, previous_mask)

compute_metrics(x: Tensor, stage: str = 'training') → dict[str, Tensor]

compute_output_shape(*args, **kwargs)

compute_output_spec(*args, **kwargs)

count_params()

Count the total number of scalars composing the weights.

Returns:

An integer count.

property dtype

Alias of layer.variable_dtype.

property dtype_policy

classmethod from_config(config)

Creates an operation from its config.

This method is the reverse of get_config, capable of instantiating the same operation from the config dictionary.

Note: If you override this method, you might receive a serialized dtype config, which is a dict. You can deserialize it as follows:

if "dtype" in config and isinstance(config["dtype"], dict):
    policy = dtype_policies.deserialize(config["dtype"])

Args:

config: A Python dictionary, typically the output of get_config.

Returns:

An operation instance.

get_build_config()

Returns a dictionary with the layer’s input shape.

This method returns a config dict that can be used by build_from_config(config) to create all states (e.g. Variables and Lookup tables) needed by the layer.

By default, the config only contains the input shape that the layer was built with. If you’re writing a custom layer that creates state in an unusual way, you should override this method to make sure this state is already created when Keras attempts to load its value upon model loading.

Returns:

A dict containing the input shape associated with the layer.

get_config()

Returns the config of the object.

An object config is a Python dictionary (serializable) containing the information needed to re-instantiate it.

get_weights()

Return the values of layer.weights as a list of NumPy arrays.

property input

Retrieves the input tensor(s) of a symbolic operation.

Only returns the tensor(s) corresponding to the first time the operation was called.

Returns:

Input tensor or list of input tensors.

property input_dtype

The dtype layer inputs should be converted to.

property input_spec

load_own_variables(store)

Loads the state of the layer.

You can override this method to take full control of how the state of the layer is loaded upon calling keras.models.load_model().

Args:

store: Dict from which the state of the model will be loaded.

property losses

List of scalar losses from add_loss, regularizers and sublayers.

property metrics

List of all metrics.

property metrics_variables

List of all metric variables.

property non_trainable_variables

List of all non-trainable layer state.

This extends layer.non_trainable_weights to include all state used by the layer including state for metrics and `SeedGenerator`s.

property non_trainable_weights

List of all non-trainable weight variables of the layer.

These are the weights that should not be updated by the optimizer during training. Unlike, layer.non_trainable_variables this excludes metric state and random seeds.

property output

Retrieves the output tensor(s) of a layer.

Only returns the tensor(s) corresponding to the first time the operation was called.

Returns:

Output tensor or list of output tensors.

property path

The path of the layer.

If the layer has not been built yet, it will be None.

property quantization_mode

The quantization mode of this layer, None if not quantized.

quantize(mode, type_check=True)

quantized_build(input_shape, mode)

quantized_call(*args, **kwargs)

rematerialized_call(layer_call, *args, **kwargs)

Enable rematerialization dynamically for layer’s call method.

Args:

layer_call: The original call method of a layer.

Returns:

Rematerialized layer’s call method.

save_own_variables(store)

Saves the state of the layer.

You can override this method to take full control of how the state of the layer is saved upon calling model.save().

Args:

store: Dict where the state of the model will be saved.

set_weights(weights)

Sets the values of layer.weights from a list of NumPy arrays.

stateless_call(trainable_variables, non_trainable_variables, *args, return_losses=False, **kwargs)

Call the layer without any side effects.

Args:

trainable_variables: List of trainable variables of the model. non_trainable_variables: List of non-trainable variables of the

model.

*args: Positional arguments to be passed to call(). return_losses: If True, stateless_call() will return the list of

losses created during call() as part of its return values.

**kwargs: Keyword arguments to be passed to call().

Returns:

A tuple. By default, returns (outputs, non_trainable_variables).

If return_losses = True, then returns (outputs, non_trainable_variables, losses).

Note: non_trainable_variables include not only non-trainable weights such as BatchNormalization statistics, but also RNG seed state (if there are any random operations part of the layer, such as dropout), and Metric state (if there are any metrics attached to the layer). These are all elements of state of the layer.

Example:

model = ...
data = ...
trainable_variables = model.trainable_variables
non_trainable_variables = model.non_trainable_variables
# Call the model with zero side effects
outputs, non_trainable_variables = model.stateless_call(
    trainable_variables,
    non_trainable_variables,
    data,
)
# Attach the updated state to the model
# (until you do this, the model is still in its pre-call state).
for ref_var, value in zip(
    model.non_trainable_variables, non_trainable_variables
):
    ref_var.assign(value)

property supports_masking

Whether this layer supports computing a mask using compute_mask.

symbolic_call(*args, **kwargs)

property trainable

Settable boolean, whether this layer should be trainable or not.

property trainable_variables

List of all trainable layer state.

This is equivalent to layer.trainable_weights.

property trainable_weights

List of all trainable weight variables of the layer.

These are the weights that get updated by the optimizer during training.

property variable_dtype

The dtype of the state (weights) of the layer.

property variables

List of all layer state, including random seeds.

This extends layer.weights to include all state used by the layer including `SeedGenerator`s.

Note that metrics variables are not included here, use metrics_variables to visit all the metric variables.

property weights

List of all weight variables of the layer.

Unlike, layer.variables this excludes metric state and random seeds.