Ramanujan

Currently, it is only a GPT-2 124M.

Small Language Model for specific usage.

Developer

git clone --recurse-submodules https://github.com/SkywardAI/ramanujan.git

Pull the code of the submodule

git submodule update --init --recursive

Check code quality

make linter

Acknowledgment

• https://github.com/SkywardAI/bundoora
• https://youtu.be/7xTGNNLPyMI?feature=shared
• https://www.kaggle.com/code/aisuko/implement-gpt2-124m-and-load-hf-gpt2