6. Batch Normalization

Medium

Implement the batch normalization forward pass.

Signature: def batchnorm_forward(x: np.ndarray, gamma: np.ndarray, beta: np.ndarray, eps: float = 1e-5) -> np.ndarray

Normalize across the batch dimension (axis=0), then scale by gamma and shift by beta.

Example:

Math

\overset{x}{^} = \frac{x - μ}{σ ^{2} + ϵ}, y = γ \overset{x}{^} + β

Asked at

Python 30/10 runs today

Output

Anything you print() in your code will show up here after you click Run.

Test Results

○identity scale/shift

○scaled output

○three rows🔒 Premium

○shift-invariant: bn(x) == bn(x + c) (mean subtracted)

○output is centered: column means ≈ 0 with identity gamma/beta