Pipeline Templates
The Template Gallery provides ready-made pipelines for common machine learning tasks. Each template is pre-configured with nodes, connections, and sample datasets so you can start training immediately.
Opening the Template Gallery
Section titled “Opening the Template Gallery”Click Templates in the toolbar to open the gallery. You can also use keyboard shortcut number keys (1–4) inside the gallery to switch categories.
Available Templates
Section titled “Available Templates”Classification
Section titled “Classification”| Template | Difficulty | Dataset | Description |
|---|---|---|---|
| Basic Classification | Beginner | Iris | Random Forest classifier with evaluation |
| Binary Classification | Beginner | Titanic | Logistic Regression with train/test split and stratification |
| Tuned Classification | Intermediate | Iris | Random Forest with Optuna hyperparameter tuning |
Regression
Section titled “Regression”| Template | Difficulty | Dataset | Description |
|---|---|---|---|
| Basic Regression | Beginner | California Housing | Linear Regression with evaluation |
| Random Forest Regression | Intermediate | California Housing | Random Forest with data splitting |
| Gradient Boosting Regression | Intermediate | Wine Quality | Gradient Boosting ensemble method |
Advanced
Section titled “Advanced”| Template | Difficulty | Dataset | Description |
|---|---|---|---|
| Full ML Pipeline | Advanced | Iris | Load → Split → Train → Evaluate → Export |
| Custom Preprocessing | Intermediate | Titanic | Script node for custom Python preprocessing |
Loading a Template
Section titled “Loading a Template”- Open the Template Gallery
- Browse or filter by category (All, Classification, Regression, Advanced)
- Click on a template card to load it
What Templates Include
Section titled “What Templates Include”Each template sets up:
- Nodes — Pre-configured with appropriate settings (model type, target column, split ratio)
- Connections — Nodes are wired together in the correct order
- Sample data — Bundled datasets (Iris, Titanic, California Housing, Wine Quality) are loaded automatically
Templates do not include saved runs, experiments, or trained models. After loading, click Run to execute the pipeline.
Customizing a Template
Section titled “Customizing a Template”After loading a template, you can:
- Change the dataset — Click the DataLoader node and browse for your own CSV file
- Switch algorithms — Click the Trainer node and select a different model type
- Add nodes — Drag new nodes from the toolbar and connect them
- Adjust parameters — Modify train/test split ratio, hyperparameters, etc.
Templates and Save/Load
Section titled “Templates and Save/Load”- Templates are built-in and cannot be modified or deleted
- Once loaded, the pipeline behaves like any other — you can save it with a custom name
- Saving a pipeline that was loaded from a template creates an independent copy
Keyboard Navigation
Section titled “Keyboard Navigation”Inside the gallery dialog:
| Key | Action |
|---|---|
| Arrow keys | Navigate between template cards |
| Enter | Load the focused template |
1–4 | Switch category tabs |
| Escape | Close the gallery |