braindecode.visualization.deep_lift#

braindecode.visualization.deep_lift(model, x, target, baseline=None)[source]#

DeepLIFT attribution (Shrikumar et al., 2017) via captum.

baseline defaults to zeros. Pass a tensor with the same shape as x to use a different reference (e.g. a per-channel mean).

Examples using `braindecode.visualization.deep_lift`#