batch_norm

dragon.vm.torch.nn.functional.batch_norm(
  input,
  running_mean,
  running_var,
  weight,
  bias,
  training=False,
  momentum=0.1,
  eps=1e-05
)[source]

Apply the batch normalization to input. [Ioffe & Szegedy, 2015].

The normalization is defined as:

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

The moving average of stats are calculated as:

\[x_{moving} \leftarrow (1 - momentum) * x_{moving} + momentum * x_{stat} \]
Parameters:
Returns:

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