1 unstable release
0.1.0 | Sep 19, 2024 |
---|
#186 in Machine learning
27KB
365 lines
micrograd_rs
A Rust Beginner's implementation of Andrej karpathy's micrograd.
Here's Link to the YouTube video.
Also thanks to micrograd-rust for reference implementation in rust.
The current implementation is limited to:
==============================================
- scalar valued autograd
- Implementation of forward and backward pass for arithmetic operations and power operation.
- Implementation of forward and backward pass for activation functions - tanh and relu.
Sample Output:
===============================================
TO-DO:
==========================================
- graph visulization using libraries like graphviz
- Implementation of Tensor type library from scratch
- Improve implementation of MLP, activations etc for Tensors.
Dependencies
~0.7–1.2MB
~25K SLoC