You can even start from scratch - micrograd walktrhough

see also

• Building a neural network FROM SCRATCH
• Fast LLM Inference From Scratch - Pushing single-GPU inference throughput to the edge without libraries

micrograd / youtube

From previous Sr. Director of AI at Tesla (and Rubik’s cube color extractor author)

Build as a tutorial to learn backpropagation & NN principle.

• micrograd from scratch
• MicroGradCpp / github - MicroGrad implementation in C++ with a simple API.
  • micrograd-cpp
• ML – Neural Network Implementation in C++ From Scratch

Python

• Tinygrad: A simple and powerful neural network framework - 1000 loc python

• Thinc - python
• nnabla - python by Sony

C / C++

• LibNC - C Library for Tensor Manipulation by Fabrice Bellard (binary only / code on demand)
• Dlib - c++