L2

tflearn.losses.L2 (tensor, wd=0.001)

Computes half the L2 norm of a tensor without the sqrt:

output = sum(t ** 2) / 2 * wd

Arguments

  • tensor: Tensor. The tensor to apply regularization.
  • wd: float. The decay.

Returns

The regularization Tensor.


L1

tflearn.losses.L1 (tensor, wd=0.001)

Computes the L1 norm of a tensor:

output = sum(|t|) * wd

Arguments

  • tensor: Tensor. The tensor to apply regularization.
  • wd: float. The decay.

Returns

The regularization Tensor.