How kareemmas3ud/chebyshev-ai-matcher works
This project generates Chebyshev transformer training data, trains an MLP to predict section impedances, and provides a Streamlit interface for user inputs, predictions, theoretical comparison, and frequency-response visualization. The README documents MATLAB dataset generation; the sampled app source shows it trains its own model from the Al Model spreadsheet rather than loading a model produced by train_model.py. The available evidence does not establish how the two spreadsheet copies are synchronized.
Main entry points: app.py Streamlit Interface · chebychev_Transformer.m Dataset Generator · train_model.py CLI Model Training
Components 12
Dataset Generation
- Dataset Generator
MatLab Code/chebychev_Transformer.m - Generated Dataset
- Training Dataset
Model Training
- CLI Model Training
Al Model/train_model.py - In-App Training
Al Model/app.py - Impedance Prediction
Al Model/app.py
Design Application
- Streamlit Interface
Al Model/app.py - Design Parameters
Al Model/app.py
Validation and Analysis
- Theory Calculation
Al Model/app.py - Error Comparison
Al Model/app.py - Frequency Response
Al Model/app.py
Other components
- Design User
Connections 14
- Dataset Generator → Generated Dataset: writes
- Training Dataset → CLI Model Training: loads
- Training Dataset → In-App Training: loads
- Design User → Streamlit Interface: uses
- Streamlit Interface → Design Parameters: collects
- Streamlit Interface → In-App Training: trains
- Streamlit Interface → Impedance Prediction: predicts
- Design Parameters → Impedance Prediction: provides inputs
- Design Parameters → Theory Calculation: provides inputs
- Impedance Prediction → Error Comparison: compares
- Theory Calculation → Error Comparison: compares
- Streamlit Interface → Frequency Response: plots
- Design Parameters → Frequency Response: provides load
- Theory Calculation → Frequency Response: provides constant