Creating Complete Dataset

This guide explains how to create a complete dataset in WrenchML after you've configured and validated your defect parameters through sample images.

Overview of Dataset Creation

Creating a complete dataset is the process of generating a large set of synthetic images with corresponding annotations that you can use to train computer vision models. This step follows after you've:

1. Uploaded your CAD model
2. Configured defect parameters
3. Generated and approved sample images

The Dataset Creation Process

Step 1: Finalize Your Configuration

Before creating a complete dataset, ensure you're satisfied with:

• The physical properties (material, color, lens type, background)
• The defect configuration (defect types, weights, and specific parameters)
• The sample images generated with these settings

<aside> Note: Once a dataset is created, its configuration cannot be changed. You would need to create a new dataset to use different parameters.

</aside>

Step 2: Initiate Dataset Creation

1. On the dataset configuration page (after reviewing your sample images)
2. Click the "Create Dataset" button
3. The system will begin generating your complete dataset
4. You'll be redirected to the Datasets page