Follow the Example

The following is an explanation of a usage scenario to help you understand the functions of Link. First, install and execute Link. Second, Load the sample source provided and follow the steps below to learn the key functions of Link.

📘

To run the sample source, the package listed below must me installed. Please ensure that the package is installed to avoid any errors.

• scikit-learn (install)
• pytorch-lightning (install)
• livelossplot (install)
• torch (install)

1. Install and execute Link
   
2. Load the downloaded sample source (Link_SampleCode(iris_classification).ipynb)
   
3. Open the component registration menu (in the top right corner, click 'Add to Pipeline')
   
4. Input component information to be registered
   
5. Component has been added to pipeline panel
   
6. Execute component(Click the newly added component and click the Run button in the center.)
   
7. Execute data cell
   
8. Create a data frame and connect the parent component
   
9. Component connection complete
   
10. Create Train, Valid split components, and connect them to the parent component
    : Split data into training and validation datasets.
    
11. Component connection complete
    
12. Create a data loader component and connect it to the parent component
    
13. Component connection complete
    
14. Define Model component creation (No parent component connected)
    : Model definition and creation are independent and unrelated to data loading, therefore they are not linked to the parent component.
    
    
15. Define Model component creation complete
    
16. Define Model component creation and connect it to the parent component
    
17. Create Model component creation and connect parent component
    
18. Define Logger component creation (No connection to parent component)
    : Define logger to determine whether the model has properly learned.
    
    
19. Complete the Define logger component creation
    
20. Connect parent component once train model is created.
    : Connect 3 parent components (data loader, create model, define logger) for model training.
    
21. Train model component creation complete
    
22. Train Model execution
    : Execute the Train Model component to run the connected components in sequence and inspect the results.