Implement and vectorize NumPy Conv2D
TeslaMachine Learning Engineer · Technical Screen
Last updated
This question evaluates understanding of 2D convolution mechanics, multidimensional NumPy array manipulation, and the competency to optimize numerical computations for performance and memory.
Implement a 2D convolution operation from scratch using NumPy only (no TensorFlow or PyTorch). Assume NCHW input shape (N, C_in, H_in, W_in) and weights of shape (C_out, C_in, k_h, k_w); support configurable stride and padding. First provide a clear nested-loop reference implementation; then optimize by vectorizing the computation (e.g., im2col or stride tricks) and discuss time/memory trade-offs.
Solution walkthrough is for members
Get the full walkthrough with reference code and complexity analysis, plus Kitty Coach at 300 messages a day. Hints, the code runner and community answers stay free.