Tesla

NumPy Conv2D Forward and Parameter Count

Members only

Implement a Conv2D forward pass in NumPy without using deep-learning framework kernels. The round can also ask for output-shape reasoning, total parameter count, bias handling, and vectorization.