ReLU

class gurobi_ml.modeling.neuralnet.activations.ReLU

Bases: object

Class to apply the ReLU activation on a neural network layer.

Parameters:

• setbounds (Bool) – Optional flag not to set bounds on the output variables.

• bigm (Float) – Optional maximal value for bounds use in the formulation

setbounds

Optional flag not to set bounds on the output variables.

Type:

Bool

bigm

Optional maximal value for bounds use in the formulation

Type:

Float

Methods

mip_model(layer)

MIP model for ReLU activation on a layer.

mip_model(layer)

MIP model for ReLU activation on a layer.

Parameters:

layer (AbstractNNLayer) – Layer to which activation is applied.