group_norm¶

dragon.vm.torch.nn.functional.group_norm(
  input,
  weight,
  bias,
  groups=32,
  eps=1e-05
)[source]¶

Apply the group normalization to input. [Wu & He, 2018].

The normalization is defined as:

\[y = \frac{x - \mathrm{E}[x]}{\sqrt{\mathrm{Var}[x] + \epsilon}} * \gamma + \beta \]

Parameters:

• input (dragon.vm.torch.Tensor) – The input tensor.
• weight (dragon.vm.torch.Tensor) – The weight tensor.
• bias (dragon.vm.torch.Tensor) – The bias tensor.
• groups (int, optional, default=32) – The number of groups to split.
• eps (float, optional, default=1e-5) – The epsilon value.

Returns:

dragon.vm.torch.Tensor – The output tensor.