Python class

GroupNorm

`GroupNorm`

class max.nn.GroupNorm(num_groups, num_channels, eps=1e-05, affine=True, device=gpu:0)

source

Bases: Module

Group normalization block.

This layer divides channels into groups and computes normalization statistics per group.

When called, GroupNorm accepts a TensorValue of shape (N, C, *) where C is the number of channels. Then, it returns a normalized TensorValue of the same shape.

Parameters:

num_groups (int) – The number of groups to divide the channels into.
num_channels (int) – The number of input channels.
eps (float) – A small value added to the denominator for numerical stability.
affine (bool) – Whether to apply a learnable affine transformation after normalization.
device (DeviceRef) – The target DeviceRef for computation.

GroupNorm​

`GroupNorm`