`autopycoin.layers`.BaseBlock#

class autopycoin.layers.BaseBlock(*args, **kwargs)[source]#

Base class of a nbeats block.

Your custom block needs to inherit from it.

Parameters

label_widthint: Horizon time to forecast.
n_neuronsint: Number of neurons in the fully connected layers.
drop_ratefloat: Rate of the dropout layer. This is used to estimate the epistemic error. Expected a value between 0 and 1. Default to 0.
namestr: The name of the layer. It defines also the block_type attribute.

Raises

ValueError: If name doesn’t contain Block. drop_rate is not between 0 and 1. All others arguments are not strictly positive integers.

Attributes

label_widthint: Return the label_width.
input_widthint: Return the input_width.
input_specÌnputSpec: InputSpec instance(s) describing the input format for this layer.
drop_ratefloat: Return the drop rate.
is_interpretablebool: Return True if the block is interpretable.
is_g_trainablebool: Return True if the last layer is trainable.
block_typestr: Return the block type.

abstract coefficient_factory(*args: list, **kwargs: dict) → tensorflow.python.framework.ops.Tensor[source]#

Create the coefficients used in the last layer a.k.a g constrained layer. This method needs to be overriden.

Raises

abstract get_coefficients(output_last_dim: int, branch_name: str) → tensorflow.python.framework.ops.Tensor[source]#

Return the coefficients used in the forecast and backcast layer a.k.a g layer by calling coefficient_factory. This method needs to be overriden.

Raises

NotImplementedError: If not overriden
ValueError: Raise an error if the coefficients tensor shape is not equal to (d0, …, output_first_dim_forecast, label_width).

property is_g_trainable: bool#: Return True if the last layer is trainable.

property is_interpretable: bool#: Return True if the block is interpretable.

autopycoin.layers.BaseBlock#