https://github.com/alexbuccheri/custom_gpt2

Implementation of GPT2, as described by [Andrej Karpathy](https://www.youtube.com/watch?v=kCc8FmEb1nY&t=638s)
https://github.com/alexbuccheri/custom_gpt2

Last synced: about 1 month ago
JSON representation

Implementation of GPT2, as described by [Andrej Karpathy](https://www.youtube.com/watch?v=kCc8FmEb1nY&t=638s)

• Host: GitHub
• URL: https://github.com/alexbuccheri/custom_gpt2
• Owner: AlexBuccheri
• Created: 2024-01-08T17:29:46.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-01-25T20:44:58.000Z (12 months ago)
• Last Synced: 2024-10-28T10:28:29.334Z (3 months ago)
• Language: Python
• Size: 429 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Custom GPT2 Implementation

Learning project, following the Youtube video of

[Andrej Karpathy](https://www.youtube.com/watch?v=kCc8FmEb1nY&t=638s), where he implements GPT2 from

scratch, using Pytorch.

## Installation

* Set up a Conda environment that contains pytorch (not documented here).

* Modify `CONDA_ENV_NAME` in [makefile.common](makefile.common) to equal the name of the Conda environment.

* Run `make install-dev` in the project root.

### Testing

* Run pytest for the local conda env by typing `pytest` in the project root.

* Run `tox` for python versions 3.10 - 3.12 by typing `make test` in the project root.