Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/teddyrendahl/makesmores
Rust Implementation of Andrej Karpathy's makemore Library
https://github.com/teddyrendahl/makesmores
Last synced: 11 days ago
JSON representation
Rust Implementation of Andrej Karpathy's makemore Library
- Host: GitHub
- URL: https://github.com/teddyrendahl/makesmores
- Owner: teddyrendahl
- Created: 2024-06-01T15:50:50.000Z (7 months ago)
- Default Branch: master
- Last Pushed: 2024-06-28T00:52:33.000Z (6 months ago)
- Last Synced: 2024-06-28T02:13:46.622Z (6 months ago)
- Language: Rust
- Size: 119 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# makesmores
Rust implementation of Andrej Karpathy's [makemore](https://youtu.be/PaCmpygFfXo?si=7Zz_Qq9WKmcVCC_c) YouTube series`makesmores` takes a database of strings (in this case the 32k most popular baby names from [ssa.gov](https://www.ssa.gov/oact/babynames/) for the year 2018) and generates more of them.
Currently allows for the choice between two different model architectures:
* MLP, following [Bengio et al. 2003](https://www.jmlr.org/papers/volume3/bengio03a/bengio03a.pdf)
* CNN, following [DeepMind WaveNet 2016](https://arxiv.org/abs/1609.03499)This is intended as a low-level exploration of neural networks, so although it depends on `torch`, it only utilizies low level `Tensor` operations. Layers of the network are implemented by hand. Hyper parameters can be adjusted via the constants at the top of the file.
Although this code represents the product at the end of the video series, for those following along, the partial implementations can be found in the commit history.