Batch Normalization Forward

Apply training-mode batch normalisation to a 2-D batch X of shape (batch, features): for each feature compute the mean and population variance over the batch, normalise with $\sqrt{\text{var}+\epsilon }\backslash sqrt\{\backslash text\{var\}+\backslash varepsilon\}$, then apply the per-feature scale gamma and shift beta. eps defaults to ${10}^{-5}10^\{-5\}$.

Implement solve(X, gamma, beta, eps=1e-5).

Returns. Return a float NumPy array with the shape of X.

Examples

Example 1

Input:

solve([[1, 2], [3, 4]], [1, 2], [0, 1], eps=0)

Output:

[[-1.0, -1.0], [1.0, 3.0]]

Example 2

Input:

solve([[2], [2]], [3], [4])

Output:

[[4.0], [4.0]]