- Setup Github account
- Install required packages:
pip install -r requirements.txt
- Create toy application with Streamlit.
- Push repository to GitHub.
- Deploy on Streamlit community cloud.
Sample application code: toy-app.py
- Perform EDA and model development on Jupyter notebook.
- Develop a training script to automate model training and persistance.
- Run the training script to train a loan approval model:
python src/training.py --data_path data/loan_dataset.csv --model_path models/ --f1_criteria 0.6
Sample model training notebook: DSSI_LoanModel.ipynb
Sample training script: training.py
- Develop an inference script to serve predictions.
- Create a loan approval application with Streamlit that automates decisions with user inputs and trained model.
Sample application code: app.py
Sample inference script: inference.py
Run and test the application locally:
streamlit run app.py
- Commit repository to GitHub.
- Deploy on Streamlit community cloud.# dssi-py_new