reference/generated/torchrl.modules.SimplicialNormalization

Run in Google Colab

Colab

Download Notebook

Notebook

View on GitHub

GitHub

SimplicialNormalization

class torchrl.modules.SimplicialNormalization(dim: int)

Apply softmax independently to fixed-size feature simplices.

Parameters:

dim (int) – Number of features in each simplex. The last input dimension must be divisible by dim.

forward(x: Tensor) → Tensor

Normalize the last dimension of x in groups of dim.