Adding KantanLibrary
KantanLibrary is a repository of stock datasets in various domains that can be used as complementary training data for your engines.
This video will show you how to add KantanLibrary datasets to your engine with the purpose of improving its scores and word counts.
1. Click the relevant engine from the Dashboard to make it active.
2. Navigate to the Training tab and click ‘Library’.
3. In the pop-up dialogue box, you will see all the available KantanLibrary datasets sorted by language and domain. If you hover your mouse over them, you will see a tool tip displaying the source and target word counts. If you select several datasets, their wordcount will be added to the total. Choose the most convenient for you and click ‘Select’.
4. After adding all the data you need, click Build. This will train your engine with both KantanLibrary and your own training data.
5. You can track the status of the build in Jobs.
6. When the build is complete, you can check the engine scores at the bottom of the Dashboard or in BuildAnalytics.
If you encounter any problems or need help, please contact us at [email protected]