Udacity MLOPS nanodegree project to train a model on census data and publish it via API on Render
Github Link for Project: https://github.com/adityajn105/udacity-mlops/
- Have python3.8 installed.
- Firstly install all requirements:
pip install -r requirements.txt
- Build models using
python mlpipeline.py
- You can test using
pytest
EDA notebook is provided: link. This is referred while developing cleaning procedures.
To clean the data, it will create cleaned_census.csv in data folder.
python mlpipeline.py --step=clean
To train model, it will create necessary model files in model folder.
python mlpipeline.py --step=train_save_model
Evaluate model on different slices, output can be found in data folder.
python mlpipeline.py --step=evaluate
python mlpipeline.py
python check_render_api.py
This project is supported by CI/CD, every new commit to main
or pull request to main
will trigger a github action workflow.
Link to CI pipeline