Label automatically with models
V7's Models feature opens up the ability to use a trained model to automatically label your data. In this guide, we'll take a look at how you can structure your workflow to take the manual work out of labelling.
Getting started
The crucial ingredient to our workflow here will be our model. If you haven't trained a model in V7 yet, follow the steps here to get one running.
If you've already created a model, head to Models, click Start, and select the minimum and maximum number of servers to use when running your model. This will determine the speed with which your model is able to label items.
You can also choose to Start when invoked - this will automatically turn you model on as soon as an inference request is made. In other words, as soon as a file is sent through the model stage in your workflow, the model will turn on automatically.
Stop when idle will turn off your model if no requests have been made in the past hour to prevent the model from running in the background if it is not being used.
Set up your workflow
Now that your model is ready to use, head to the Workflows page and hit Create Workflow to open up the Workflow creation wizard.
You can also access the workflow creation wizard in the final stage of creating any new dataset.
Add a Model stage, click Connect Model, and map the classes the model was trained on with the corresponding classes that the model will output.
Select the option to Auto Start. This will turn on the model as soon as any items are sent through the model stage.
Label
When Auto-Start is enabled, annotators can trigger the model by simply sending items to the next stage when a model stage follows an Annotate or Review stage.
Workforce Managers, Users, and Admins can also send files to the Model stage in bulk by filtering for all items in the stage prior, selecting all and updating their stage.
Once the Model stage has labelled your data, your reviewers can step in to make any required changes to the model's inference.
Updated 6 months ago