spconv ... sparse convolution

### Motivation and description

For many applications (graph neural networks, 3d point clouds, ...) sparse convolution would be a cool feature

### Possible Implementation

A guide to implementation is given here (with some references):
https://towardsdatascience.com/how-does-sparse-convolution-work-3257a0a8fd1

Begin with the motivation and the linked Towards Data Science implementation guide. The issue names no files, entry points, or tests; before coding, establish the intended API and supported sparse-convolution cases, with completion defined by an accepted implementation and tests.

Written by the indexing model from the issue text.