AI Base:

Tools for AI Engineering and Machine Learning

Overview

This project contains a curated set of Python tools and libraries that one would need for work in AI Engineering and Machine Learning such as jupyterLab, langchain, tensorflow, pytorch and transformers. To see the full set of what's included, see the requirements.txt file.

These tools have been tested using Python 3.12, the latest version currently supported by tensorflow. The versions are currently unpinned, however, so migrating to different versions should be straightforward to the extent that the tools are supported.

Installation and Usage

These tools can be used with either pip, conda, or docker, depending on your preferences.

Using Pip

As always, using a virtual environment is strongly recommended. With your environment activated, simply use:

pip install -r requirements.txt

Using Conda

To create the environment from scratch, in the root directory:

conda env create -f environment.yml

Activate and use with:

conda activate aibase

Using Docker

There's a docker image, codesolid/ailatest that you can use as a starting point for your AI projects.

If you have the task task runner installed, you can use it to run the tasks in Taskfile.yml.

If you prefer not to install that, here is a summary:

Pulling the image

docker pull codesolid/aibase
docker tag codesolid/aibase aibase

Run a bash shell using the image

docker run -p 8888:8888 -it --rm --env-file .env -v ${PWD}:/app aibase bash

Run Jupyter Notebook

docker run -u jupyter -p 8888:8888 -it --rm --env-file .env -v ${PWD}:/app aibase jupyter lab --no-browser --ip="0.0.0.0"  --NotebookApp.token='' --NotebookApp.password=''

Building the image:

From within the source directory, run:

- docker build -t aibase .