This is the repo for the Gazella project, which aims to build and share nano GPT-style models.
You'll find in this repo:
src/gazella/
- source code for modelsscripts/
- scripts to run LLM workloadsdata/
- prepare datatrain/
- train or finetune modelsinference/
- convert models to other format, and generate responses
Gazelles are known as swift animals.
Gazella is released under the Apache License 2.0. Read the LICENSE file for more details.