pub fn batch_norm<const D: usize>(
input: Tensor<D>,
gamma: Tensor<1>,
beta: Tensor<1>,
mean: Tensor<1>,
variance: Tensor<1>,
epsilon: f64,
) -> Tensor<D>Expand description
Applies batch normalization using explicitly supplied channel statistics.
input has shape [batch, channels, ...]; gamma, beta, mean, and
variance each have shape [channels].
This function doesn’t calculate or update statistics. Callers may supply running statistics for inference or batch statistics calculated by a training path.