Baremetal-NN
Baremetal-NN API documentation
|
void NN_batch_norm2d | ( | Tensor * | out, |
const Tensor * | in, | ||
const Tensor * | weight, | ||
const Tensor * | bias, | ||
float | eps, | ||
const Tensor * | running_mean, | ||
const Tensor * | running_va | ||
) |
Applies Batch Normalization over a 4D input.
out | the output tensor of shape (batch_size, height, width, channels) |
in | the input tensor of shape (batch_size, height, width, channels) |
weight | the learnable weights of the module of shape (channels), or NULL if no weight is applied |
bias | the learnable bias of the module of shape (channels), or NULL if no bias is applied |
eps | a value added to the denominator for numerical stability |
running_mean | the running mean of the module of shape (channels), or NULL if no running mean is applied |
running_var | the running variance of the module of shape (channels), or NULL if no running variance is applied |