Batch Normalization for BCHW Input
Initialization, Normalization & Regularization DS practice problem on Onlearn.
Difficulty: easy.
Topics: Understanding Implement Batch Normalization for BCHW Input, Batch Normalization, BCHW Data Layout, Running Statistics, Affine Transformation, Epsilon Smoothing, Deep Learning, Numerical Computing, Computer Vision, Statistical Analysis, Tensor Algebra, Normalization Techniques, Convolutional Neural Networks, Array Manipulation, Parameter Optimization, Numerical Stability.
Implement a function that performs Batch Normalization on a 4D NumPy array representing a batch of feature maps in the BCHW format (batch, channels, height, width). The function should normalize the input across the batch and spatial dimensions for each channel, then apply scale (gamma) and shift (beta) parameters. Use the provided epsilon value to ensure numerical stability.