Pad the number of samples. Parameters ---------- start_padding : int The number of padded sa

This is in essence a DN3 wrapper for the pytorch function `interpolate() <https://pytorch.org/docs/stable/nn.functional.html>`_

Crop to a new length of `cropped_length` from the specified `start_offset`, or randomly select an offset. Parameters -------

Creates a Deep10-10 mapping for the provided dataset. Parameters ---------- dataset : Dataset max_scale : f

Transforms incoming Deep1010 data into exclusively the more limited 1020 channel set.

Uniformly selects a transform from the `transform_list` with probabilities according to `weights`. Parameters ----------

This DenseNet-inspired filter block features in the TIDNet network from Kostas & Rudzicz 2020 (Thinker Invariance). 2D convolution is

This extends the :any:`DenseFilter` to specifically operate in channel space and collapse this dimension over the course of `depth` l

This implements the dilated temporal-only spanning convolution from TIDNet. Parameters ---------- channels f

Instead of summarizing the entire temporal dimension into a single prediction, a prediction kernel is swept over the final sequence r

Initialization of the Base Trainable object. Any learning procedure that leverages DN3atasets should subclass this base class.

Add custom code to load a specific thinker from a set of session files. Warnings ---------- For all intents and purp

Add a transformation that is applied to every fetched item in the dataset Parameters ---------- transform : BaseTran

Returns ------- channels: list The channel sets used by the dataset.

Remove all added transforms from dataset.

Maps the labels returned by this to the events as recorded in the original annotations or stim channel. Returns -------