Computes mean and std for batch then apply batch_normalization on batch.
Computes mean and std for batch then apply batch_normalization on batch.
k_normalize_batch_in_training(x, gamma, beta, reduction_axes, epsilon = 0.001)
x |
Input tensor or variable. |
gamma |
Tensor by which to scale the input. |
beta |
Tensor with which to center the input. |
reduction_axes |
iterable of integers, axes over which to normalize. |
epsilon |
Fuzz factor. |
A list length of 3, (normalized_tensor, mean, variance).
This function is part of a set of Keras backend functions that enable lower level access to the core operations of the backend tensor engine (e.g. TensorFlow, CNTK, Theano, etc.).
You can see a list of all available backend functions here: https://keras.rstudio.com/articles/backend.html#backend-functions.
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.