fusionlab.nn.models.SuperXTFT

class fusionlab.nn.models.SuperXTFT

Bases: XTFT

An extension of XTFT that injects Variable Selection Networks (VSNs) and a Gate→Add&Norm→GRN refinement pipeline on attention and decoder outputs. It inherits all shared logic from BaseExtreme (validation, losses, hooks) and the baseline flow from XTFT, then overrides only the parts that differ.

Key Additions

• VariableSelectionNetwork for: - static features (no time axis), - dynamic / historical features, - future-known covariates. These learn soft weights to highlight the most informative variables at each step.

• GRN refinement blocks applied after: - Hierarchical attention output, - Cross attention output, - Memory-augmented attention output, - Decoder output (via Gate→Add&Norm→GRN pipeline).

• Optional feature-based anomaly scoring (when anomaly_detection_strategy == 'feature_based'): a dedicated attention → projection → scorer head produces per-timestep anomaly scores that feed the anomaly loss.

{key_parameters}

Design Rationale

VSNs reduce noise from high-dimensional inputs by learning which signals matter now. The extra GRNs stabilize attention outputs and improve representational depth without exploding parameters. Keeping these pieces modular lets you toggle them on variants without rewriting boilerplate.

Hook Overview

This class overrides:

• _build_components — builds VSNs, GRNs, and (optionally) anomaly-attention layers.

• _encode_inputs — runs VSNs before the usual embedding / normalization path.

• _temporal_backbone — inserts GRNs after each attention.

• _aggregate_decode — applies the Gate→Add&Norm→GRN pipeline to decoder outputs.

• _maybe_compute_anomaly_scores — computes scores for the feature-based strategy.

Everything else (compile logic, losses, serialization) is taken care of by the base classes.

Examples

Instantiate with VSNs and GRN refinements active by default:

model = SuperXTFT(
    static_input_dim=10,
    dynamic_input_dim=32,
    future_input_dim=8,
    forecast_horizon=7,
    quantiles=[0.1, 0.5, 0.9],
    multi_scale_agg='auto',
    anomaly_detection_strategy='feature_based',
)

Then compile and fit as usual:

model.compile(optimizer='adam')
model.fit([X_static, X_dynamic, X_future], y, epochs=20)

Notes

• Set the logger to DEBUG to see shapes and flow decisions.

• If you do not need feature-based anomalies, omit that strategy to skip building the extra attention head and save memory.

See also

XTFT

Baseline DRY implementation without VSN/extra GRNs.

BaseExtreme

Parent class providing shared plumbing, hooks, and logging.

__init__(*, static_input_dim, dynamic_input_dim, future_input_dim, embed_dim=32, forecast_horizon=1, quantiles=None, max_window_size=10, memory_size=100, num_heads=4, dropout_rate=0.1, output_dim=1, attention_units=32, hidden_units=64, lstm_units=64, scales=None, multi_scale_agg='auto', activation='relu', use_residuals=True, use_batch_norm=False, final_agg='last', anomaly_config=None, anomaly_detection_strategy=None, anomaly_loss_weight=1.0, architecture_config=None, fusion_mode=None, **kw)

Parameters:

• static_input_dim (int)

• dynamic_input_dim (int)

• future_input_dim (int)

• embed_dim (int)

• forecast_horizon (int)

• quantiles (str | List[float] | None)

• max_window_size (int)

• memory_size (int)

• num_heads (int)

• dropout_rate (float)

• output_dim (int)

• attention_units (int)

• hidden_units (int)

• lstm_units (int)

• scales (str | List[int] | None)

• multi_scale_agg (str | None)

• activation (str | callable)

• use_residuals (bool)

• use_batch_norm (bool)

• final_agg (str)

• anomaly_config (Dict[str, Any] | None)

• anomaly_detection_strategy (str | None)

• anomaly_loss_weight (float)

• architecture_config (Dict | None)

• fusion_mode (str | None)

• kw (Any)

Return type:

None

Methods

__init__(*, static_input_dim, ...[, ...])
add_loss(loss)	Can be called inside of the call() method to add a scalar loss.
add_metric(*args, **kwargs)
add_variable(shape, initializer[, dtype, ...])	Add a weight variable to the layer.
add_weight([shape, initializer, dtype, ...])	Add a weight variable to the layer.
build(input_shape)
build_from_config(config)	Builds the layer's states with the supplied config dict.
call(inputs[, training])
compile(optimizer[, loss])	Configures the model for training.
compile_from_config(config)	Compiles the model with the information given in config.
compiled_loss(y, y_pred[, sample_weight, ...])
compute_loss([x, y, y_pred, sample_weight, ...])	Compute the total loss, validate it, and return it.
compute_mask(inputs, previous_mask)
compute_metrics(x, y, y_pred[, sample_weight])	Update metric states and collect all metrics to be returned.
compute_output_shape(*args, **kwargs)
compute_output_spec(*args, **kwargs)
count_params()	Count the total number of scalars composing the weights.
evaluate([x, y, batch_size, verbose, ...])	Returns the loss value & metrics values for the model in test mode.
export(filepath[, format, verbose, ...])	Export the model as an artifact for inference.
fit([x, y, batch_size, epochs, verbose, ...])	Trains the model for a fixed number of epochs (dataset iterations).
from_config(config)	Creates an operation from its config.
get_build_config()	Returns a dictionary with the layer's input shape.
get_compile_config()	Returns a serialized config with information for compiling the model.
get_config()	Returns the config of the object.
get_layer([name, index])	Retrieves a layer based on either its name (unique) or index.
get_metrics_result()	Returns the model's metrics values as a dict.
get_params([deep])	Get the parameters for this learner.
get_state_tree([value_format])	Retrieves tree-like structure of model variables.
get_weights()	Return the values of layer.weights as a list of NumPy arrays.
help(**kwargs)
load(file_path[, format])	Load the learner's state from a specified file in the desired format.
load_own_variables(store)	Loads the state of the layer.
load_weights(filepath[, skip_mismatch])	Load the weights from a single file or sharded files.
loss(y, y_pred[, sample_weight])
make_predict_function([force])
make_test_function([force])
make_train_function([force])
predict(x[, batch_size, verbose, steps, ...])	Generates output predictions for the input samples.
predict_on_batch(x)	Returns predictions for a single batch of samples.
predict_step(data)
quantize(mode[, config])	Quantize the weights of the model.
quantized_build(input_shape, mode)
quantized_call(*args, **kwargs)
reconfigure(architecture_config)	Creates a new model instance with a modified architecture.
rematerialized_call(layer_call, *args, **kwargs)	Enable rematerialization dynamically for layer's call method.
reset_metrics()
save(filepath[, overwrite, zipped])	Saves a model as a .keras file.
save_own_variables(store)	Saves the state of the layer.
save_weights(filepath[, overwrite, ...])	Saves all weights to a single file or sharded files.
set_params(**params)	Set the parameters of this learner.
set_state_tree(state_tree)	Assigns values to variables of the model.
set_weights(weights)	Sets the values of layer.weights from a list of NumPy arrays.
stateless_call(trainable_variables, ...[, ...])	Call the layer without any side effects.
stateless_compute_loss(trainable_variables, ...)
summary([line_length, positions, print_fn, ...])	Prints a string summary of the network.
symbolic_call(*args, **kwargs)
test_on_batch(x[, y, sample_weight, return_dict])	Test the model on a single batch of samples.
test_step(data)
to_json(**kwargs)	Returns a JSON string containing the network configuration.
train_on_batch(x[, y, sample_weight, ...])	Runs a single gradient update on a single batch of data.
train_step(data)

Attributes

compiled_metrics
compute_dtype	The dtype of the computations performed by the layer.
distribute_reduction_method
distribute_strategy
dtype	Alias of layer.variable_dtype.
dtype_policy
input	Retrieves the input tensor(s) of a symbolic operation.
input_dtype	The dtype layer inputs should be converted to.
input_spec
jit_compile
layers
losses	List of scalar losses from add_loss, regularizers and sublayers.
metrics	List of all metrics.
metrics_names
metrics_variables	List of all metric variables.
my_params
non_trainable_variables	List of all non-trainable layer state.
non_trainable_weights	List of all non-trainable weight variables of the layer.
output	Retrieves the output tensor(s) of a layer.
path	The path of the layer.
quantization_mode	The quantization mode of this layer, None if not quantized.
run_eagerly
supports_masking	Whether this layer supports computing a mask using compute_mask.
trainable	Settable boolean, whether this layer should be trainable or not.
trainable_variables	List of all trainable layer state.
trainable_weights	List of all trainable weight variables of the layer.
variable_dtype	The dtype of the state (weights) of the layer.
variables	List of all layer state, including random seeds.
weights	List of all weight variables of the layer.

__init__(*, static_input_dim, dynamic_input_dim, future_input_dim, embed_dim=32, forecast_horizon=1, quantiles=None, max_window_size=10, memory_size=100, num_heads=4, dropout_rate=0.1, output_dim=1, attention_units=32, hidden_units=64, lstm_units=64, scales=None, multi_scale_agg='auto', activation='relu', use_residuals=True, use_batch_norm=False, final_agg='last', anomaly_config=None, anomaly_detection_strategy=None, anomaly_loss_weight=1.0, architecture_config=None, fusion_mode=None, **kw)

Parameters:

• static_input_dim (int)

• dynamic_input_dim (int)

• future_input_dim (int)

• embed_dim (int)

• forecast_horizon (int)

• quantiles (str | List[float] | None)

• max_window_size (int)

• memory_size (int)

• num_heads (int)

• dropout_rate (float)

• output_dim (int)

• attention_units (int)

• hidden_units (int)

• lstm_units (int)

• scales (str | List[int] | None)

• multi_scale_agg (str | None)

• activation (str | callable)

• use_residuals (bool)

• use_batch_norm (bool)

• final_agg (str)

• anomaly_config (Dict[str, Any] | None)

• anomaly_detection_strategy (str | None)

• anomaly_loss_weight (float)

• architecture_config (Dict | None)

• fusion_mode (str | None)

• kw (Any)

Return type:

None

help(**kwargs)

my_params = SuperXTFT(     static_input_dim,     dynamic_input_dim,     future_input_dim,     embed_dim=32,     forecast_horizon=1,     quantiles=None,     max_window_size=10,     memory_size=100,     num_heads=4,     dropout_rate=0.1,     output_dim=1,     attention_units=32,     hidden_units=64,     lstm_units=64,     scales=None,     multi_scale_agg='auto',     activation='relu',     use_residuals=True,     use_batch_norm=False,     final_agg='last',     anomaly_config=None,     anomaly_detection_strategy=None,     anomaly_loss_weight=1.0,     architecture_config=None,     fusion_mode=None )