merlin.models.tf.CGCBlock#
- class merlin.models.tf.CGCBlock(*args, **kwargs)[source]#
Bases:
merlin.models.tf.core.combinators.ParallelBlock
Implements the Customized Gate Control (CGC) proposed in [1].
References
[1] Tang, Hongyan, et al. “Progressive layered extraction (ple): A novel multi-task learning (mtl) model for personalized recommendations.” Fourteenth ACM Conference on Recommender Systems. 2020.
- Parameters
outputs (Union[List[str], List[PredictionTask], ParallelPredictionBlock]) –
- Names of the tasks or PredictionTask/ParallelPredictionBlock objects from
which we can extract the task names
expert_block (Union[Block, tf.keras.layers.Layer]) – Block that will be used for the experts
num_task_experts (int, optional) – Number of task-specific experts, by default 1
num_shared_experts (int, optional) – Number of shared experts for tasks, by default 1
add_shared_gate (bool, optional) – Whether to add a shared gate for this CGC block, by default False. Useful when multiple CGC blocks are stacked (e.g. in PLEBlock) As all CGC blocks except the last one should include the shared gate.
gate_block (Optional[Block], optional) – Optional block than can make the Gate mode powerful in converting the inputs into expert weights for averaging, by default None
gate_softmax_temperature (float, optional) – Temperature of the softmax used by the Gates for getting weights for the average. Temperature can be used to smooth the weights distribution, and is by default 1.0.
enable_gate_weights_metrics (bool, optional) – Enables logging the average gate weights on experts
name (Optional[str], optional) – Name of the CGC block, by default None
- __init__(outputs: Union[List[str], List[merlin.models.tf.prediction_tasks.base.PredictionTask], merlin.models.tf.prediction_tasks.base.ParallelPredictionBlock], expert_block: Union[merlin.models.tf.core.base.Block, keras.engine.base_layer.Layer], num_task_experts: int = 1, num_shared_experts: int = 1, add_shared_gate: bool = False, gate_block: Optional[merlin.models.tf.core.base.Block] = None, gate_softmax_temperature: float = 1.0, enable_gate_weights_metrics: bool = False, name: Optional[str] = None, **kwargs)[source]#
Methods
__init__
(outputs, expert_block[, ...])add_branch
(name, block)add_loss
(losses, **kwargs)Add loss tensor(s), potentially dependent on layer inputs.
add_metric
(value[, name])Adds metric tensor to the layer.
add_update
(updates)Add update op(s), potentially dependent on layer inputs.
add_variable
(*args, **kwargs)Deprecated, do NOT use! Alias for add_weight.
add_weight
([name, shape, dtype, ...])Adds a new variable to the layer.
apply_to_all
(inputs[, columns_to_filter])apply_to_branch
(branch_name, *block)as_tabular
([name])build
(input_shape)build_from_config
(config)calculate_batch_size_from_input_shapes
(...)call
(inputs, **kwargs)call_outputs
(outputs[, training])check_schema
([schema])compute_call_output_shape
(input_shape)compute_mask
(inputs[, mask])Computes an output mask tensor.
compute_output_shape
(input_shapes)compute_output_signature
(input_signature)Compute the output tensor signature of the layer based on the inputs.
connect
(*block[, block_name, context])Connect the block to other blocks sequentially.
connect_branch
(*branches[, add_rest, post, ...])Connect the block to one or multiple branches.
connect_debug_block
([append])Connect the block to a debug block.
connect_with_residual
(block[, activation])Connect the block to other blocks sequentially with a residual connection.
connect_with_shortcut
(block[, ...])Connect the block to other blocks sequentially with a shortcut connection.
copy
()count_params
()Count the total number of scalars composing the weights.
finalize_state
()Finalizes the layers state after updating layer weights.
from_config
(config[, custom_objects])from_features
(features[, pre, post, ...])Initializes a TabularLayer instance where the contents of features will be filtered out
from_layer
(layer)from_schema
(schema[, tags, allow_none])Instantiate a TabularLayer instance from a DatasetSchema.
get_build_config
()get_input_at
(node_index)Retrieves the input tensor(s) of a layer at a given node.
get_input_mask_at
(node_index)Retrieves the input mask tensor(s) of a layer at a given node.
get_input_shape_at
(node_index)Retrieves the input shape(s) of a layer at a given node.
get_item_ids_from_inputs
(inputs)get_output_at
(node_index)Retrieves the output tensor(s) of a layer at a given node.
get_output_mask_at
(node_index)Retrieves the output mask tensor(s) of a layer at a given node.
get_output_shape_at
(node_index)Retrieves the output shape(s) of a layer at a given node.
get_padding_mask_from_item_id
(inputs[, ...])get_weights
()Returns the current weights of the layer, as NumPy arrays.
parse
(*block)parse_block
(input)parse_config
(config[, custom_objects])post_call
(inputs[, transformations, ...])Method that's typically called after the forward method for post-processing.
pre_call
(inputs[, transformations])Method that's typically called before the forward method for pre-processing.
prepare
([block, post, aggregation])Transform the inputs of this block.
register_features
(feature_shapes)repeat
([num])Repeat the block num times.
repeat_in_parallel
([num, prefix, names, ...])Repeat the block num times in parallel.
repr_add
()repr_extra
()repr_ignore
()select_by_name
(name)Select a parallel block by name
select_by_names
(names)Select a list of parallel blocks by names
select_by_tag
(tags)Select layers of parallel blocks by tags.
set_aggregation
(value)- param value
set_post
(value)set_pre
(value)set_schema
([schema])set_weights
(weights)Sets the weights of the layer, from NumPy arrays.
super
()with_name_scope
(method)Decorator to automatically enter the module name scope.
Attributes
REQUIRES_SCHEMA
activity_regularizer
Optional regularizer function for the output of this layer.
aggregation
rtype: TabularAggregation, optional
compute_dtype
The dtype of the layer's computations.
context
dtype
The dtype of the layer weights.
dtype_policy
The dtype policy associated with this layer.
dynamic
Whether the layer is dynamic (eager-only); set in the constructor.
first
has_schema
inbound_nodes
Return Functional API nodes upstream of this layer.
input
Retrieves the input tensor(s) of a layer.
input_mask
Retrieves the input mask tensor(s) of a layer.
input_shape
Retrieves the input shape(s) of a layer.
input_spec
InputSpec instance(s) describing the input format for this layer.
is_input
is_tabular
layers
losses
List of losses added using the add_loss() API.
metrics
List of metrics added using the add_metric() API.
name
Name of the layer (string), set in the constructor.
name_scope
Returns a tf.name_scope instance for this class.
non_trainable_variables
non_trainable_weights
List of all non-trainable weights tracked by this layer.
outbound_nodes
Return Functional API nodes downstream of this layer.
output
Retrieves the output tensor(s) of a layer.
output_mask
Retrieves the output mask tensor(s) of a layer.
output_shape
Retrieves the output shape(s) of a layer.
parallel_dict
parallel_values
post
rtype: SequentialTabularTransformations, optional
pre
rtype: SequentialTabularTransformations, optional
registry
schema
stateful
submodules
Sequence of all sub-modules.
supports_masking
Whether this layer supports computing a mask using compute_mask.
trainable
trainable_variables
trainable_weights
List of all trainable weights tracked by this layer.
updates
variable_dtype
Alias of Layer.dtype, the dtype of the weights.
variables
Returns the list of all layer variables/weights.
weights
Returns the list of all layer variables/weights.