Machine Learning DevOps
🐍🤖 Tutorial on DevOps for Data Science - this uses Azure pipelines for the DevOps implementation
🚀 Detailed set up instructions
The detailed set up instructions including how to create your virtual environment and to get an Azure account can be found at 👉🏼 https://ml-devops-tutorial.readthedocs.io/en/latest/setup.html.
Pre requisites
Python 3.x
We will be using Python. so we recommend you get the latest version, which is 3.7
Git
Git is a version control software that records changes to a file or set of files. Git is especially helpful for software developers as it allows changes to be tracked (including who and when) when working on a project.
To download Git, go to the following link and choose the correct version for your operating system: https://git-scm.com/downloads.
GitHub
GitHub is a web-based service for version control using Git. You will need to set up an account at https://github.com. Basic GitHub accounts are free and you can now also have private repositories.
IDE supporting Python
Text editors are tools with powerful features designed to optimize writing code. There are several text editors that you can choose from.
Get VS code using 👉🏼 this link. This is your facilitator's favourite 💜 and it is worth trying.
If you decide to go for VSCode make sure to also have the Python extension installed. This will make your life so much easier (and it comes with a lot of nifty features 😎).
Setup
Basic setup
- Clone this repository. From a terminal:
git clone https://github.com/trallard/ml_devops_tutorial.git - Change into the setup directory
cd setup - Install the requirements:
pip install -r requirements pip install --upgrade azureml-sdk
Additional resources
📖 We will be using Azure Pipelines during the workshop so make sure to visit the docs before following the tutorial using 👉🏼this link 🚀🚀
Note that we also use the Python Azure Machine learning SDK 👉🏼 check the docs here
🚀 PRs and Issues are welcome
License
This repo is licensed using a CC-BY so you are free to use, remix, and share so long attribution is provided to the original author.