How gkrishna4346/predictive-maintenance-capstone works
This is a Streamlit predictive-maintenance app: a user enters six engine measurements, requests a prediction, and sees an engine-condition and maintenance recommendation. The app downloads a serialized AdaBoost model from Hugging Face, loads it, and invokes prediction directly. The source tree also has datasets and training-related notebooks/scripts, but the sampled scripts are placeholders and the README documents no further workflow, so those connections are not inferred.
Main entry points: app.py Streamlit app
Components 7
User interface
- Engine inputs
deployment/app.py - Streamlit app
deployment/app.py
Model inference
- Hugging Face Hub
- AdaBoost model
Prediction results
- Engine condition
deployment/app.py - Maintenance action
deployment/app.py
Other components
- Operator
Connections 6
- Operator → Engine inputs: enters readings
deployment/app.py - Engine inputs → Streamlit app: submits prediction
deployment/app.py - Streamlit app → Hugging Face Hub: downloads model
deployment/app.py - Streamlit app → AdaBoost model: loads and predicts
deployment/app.py - Streamlit app → Engine condition: displays result
deployment/app.py - Streamlit app → Maintenance action: displays action
deployment/app.py