Minimal deep learning library written from scratch in Python
SmallPebble
Minimal deep learning library written from scratch in Python, using NumPy/CuPy.
SmallPebble is a minimal automatic differentiation and deep learning library written from scratch in Python, using NumPy/CuPy.
The implementation is relatively small, and mainly in the file: smallpebble.py. To help understand it, check out this introduction to autodiff, which presents an autodiff framework that works in the same way as SmallPebble (except using scalars instead of NumPy arrays).
SmallPebble’s raison d’etre is to be a simplified deep learning implementation, for those who want to learn what’s under the hood of deep learning frameworks. However, because it is written in terms of vectorised NumPy/CuPy operations, it performs well enough for non-trivial models to be trained using it.
Highlights
- Relatively simple implementation.
- Can run on GPU, using CuPy.
- Various