https://github.com/gjbex/Deploying-LLMs-locally
Material for a training on AI tools
https://github.com/gjbex/Deploying-LLMs-locally
artificial-intelligence artificial-neural-networks deep-learning deployment llama llm machine-learning machine-learning-algorithms training training-materials
Last synced: about 2 months ago
JSON representation
Material for a training on AI tools
- Host: GitHub
- URL: https://github.com/gjbex/Deploying-LLMs-locally
- Owner: gjbex
- License: cc-by-4.0
- Created: 2024-07-19T14:57:56.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-07-10T11:06:05.000Z (3 months ago)
- Last Synced: 2025-07-10T16:53:11.105Z (3 months ago)
- Topics: artificial-intelligence, artificial-neural-networks, deep-learning, deployment, llama, llm, machine-learning, machine-learning-algorithms, training, training-materials
- Language: Jupyter Notebook
- Homepage: https://gjbex.github.io/Deploying-LLMs-locally/
- Size: 13.8 MB
- Stars: 13
- Watchers: 2
- Forks: 3
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Code of conduct: CODE_OF_CONDUCT.md
Awesome Lists containing this project
README
# Deploying LLMs locally
This is a repository that illustrates the use of various AI tools and
techniques and how to use them on local infrastructure such as HPC systems.
## What is it?
1. `local_LLMs.pptx`: PowerPoint presentation on running Large Language
Models (LLMs) on a local machine.
1. `source-code`: directory with the source code.
1. `models`: directory with scripts to download pre-trained models.
1. `data`: directory with scripts to download data.
1. `tools`: directory with tools to run LLMs on a local machine.
1. `docs`: directory for a web site on this training.
1. `CONTRIBUTING.md`: guidelines for contributing to this repository.
1. `LICENSE`: license information for this repository.
1. `CODE_OF_CONDUCT.md`: code of conduct for this repository and training.
## Conda environments
Since conda environments for machine learning have many dependencies, we
opted to have a separate environment for each directory in `source-code`,
rather than one environment for all code. This makes it easier to manage
dependencies, and to avoid conflicts between different packages that may
be required by different scripts.