How dhyeybuch/mlplayground works
mlplayground is a browser-based playground for exploring linear and logistic models. Users choose a regression or classification dataset, configure training parameters, run training, and inspect plots and results. The browser loads dataset metadata from the FastAPI backend; training requests are split into train/test sets, passed to the selected gradient-descent implementation, and returned with model-specific outputs. The frontend then visualizes the response and exposes generated example code. The README is brief, so the feature scope here is based mainly on the sampled frontend and backend source.
Main entry points: app.js Selection and controls · main.py Training endpoint · index.html Playground interface · datasets.py Dataset catalog
Components 12
Browser experience
- Playground interface
frontend/index.html - Selection and controls
frontend/app.js
API and data
- Dataset endpoint
backend/main.py - Dataset catalog
backend/datasets.py - Training endpoint
backend/main.py
Model training
- Train-test split
backend/main.py - Gradient descent models
backend/models.py - Model results
backend/main.py
Results display
- Plots and curves
frontend/app.js - Metrics and code
frontend/app.js
Other components
- Playground user
- Plotly
Connections 12
- Playground user → Playground interface: uses
- Playground interface → Selection and controls: hosts controls
- Selection and controls → Dataset endpoint: loads datasets
frontend/app.js - Dataset endpoint → Dataset catalog: loads catalog
backend/main.py - Selection and controls → Training endpoint: submits training
frontend/app.js - Training endpoint → Dataset catalog: finds dataset
backend/main.py - Training endpoint → Train-test split: splits data
backend/main.py - Training endpoint → Gradient descent models: trains selected model
backend/main.py - Training endpoint → Model results: builds response
backend/main.py - Selection and controls → Plots and curves: plots response
frontend/app.js - Selection and controls → Metrics and code: updates results
frontend/app.js - Plots and curves → Plotly: renders plots
frontend/app.js