You can create and manage your NLP models in the « Model » section.
The home screen of this section shows you all the models available in your project, wether trained or ready to be trained.
Click on a model to have access to the main characteristics of its training (training sentences, entities, test sentences).
TRAINING A MODEL
1. Create a new model
Click on the « New model » button to create a « New model ».
Give it a name and choose the training configuration that will be used to train this model.
2. Create your training set
By default all the sentences available in your project will be added to your training set.
The number of sentences in your training set is displayed in the left corner of the window.
You can refine your sentence selection by using the filter options or by selecting a pre-defined view.
Click on « Next » to continue.
3. Create your test sets
Once your training set configured, you have the option to create test sets. These test sets will be used to mesure the performance of your model on indicators such as precision, recall and F1-score.
Click on « add a test set » to create a new test set and use the filter option to refine your sentence selection. Multiple test sets can be added for each model.
Click on next to validate your test set and continue.
4. Entity selection
You can choose to include or not the different entities available in your project. By default, all the entities found in your training set sentences are added. Click on the checkbox to disable an entity.
Click on next to validate your model entity and continue.
Click on save to finalize the creation of your model. It is now ready to be trained and will be displayed in your model list. Click on the « Train » button to launch the training. Depending on the size of your training set, the process can take a few seconds up to several minutes to complete.