qiskit.aqua.components.neural_networks.PyTorchDiscriminator.gradient_penalty¶
-
PyTorchDiscriminator.gradient_penalty(x, lambda_=5.0, k=0.01, c=1.0)[source]¶ Compute gradient penalty for discriminator optimization
- Parameters
x (numpy.ndarray) – Generated data sample.
lambda (float) – Gradient penalty coefficient 1.
k (float) – Gradient penalty coefficient 2.
c (float) – Gradient penalty coefficient 3.
- Returns
Gradient penalty.
- Return type
torch.Tensor