Conclusion¶

Congratulations! You have successfully completed the fifth part of the guide. In this part, you learned how to annotate the data in Label Studio with AI-assisted labeling. You finally retrained the model using the newly labeled data.

As you work with your data and model, you may find that certain labels need refinement or that additional data points require annotation. You can continuously improve the quality of your labeled dataset, to improve your model accuracy and effectiveness over time. Based on the model performance feedback, you can revisit and update the annotations to create a broader and more robust training set that better meets the needs of your project.

The following diagram illustrates the bricks you set up at the end of this part:

flowchart TB
    extra -->|upload| labelStudioTasks
    labelStudioTasks -->|label| labelStudioAnnotations
    bento_model -->|load| fastapi
    labelStudioTasks -->|POST /predict| fastapi
    fastapi --> labelStudioPredictions
    labelStudioPredictions -->|submit| labelStudioAnnotations
    labelStudioAnnotations -->|download| extra_annotations
    extra_annotations -->|load| parse_annotations
    parse_annotations -->|copy| data_raw
    data_raw -->|dvc repro| bento_model

    subgraph workspaceGraph[WORKSPACE]
        extra[extra-data/extra]
        extra_annotations[extra-data/annotations.json]
        bento_model[model/classifier.bentomodel]
        fastapi[src/serve_labelstudio.py]
        parse_annotations[scripts/parse_annotations.py]
        data_raw[data/raw]
    end

    subgraph labelStudioGraph[LABEL STUDIO]
        labelStudioTasks[Tasks]
        labelStudioAnnotations[Annotations]
        labelStudioPredictions[Predictions]
    end

Going further¶

Info

Retrain on the cluster: In this part you ran dvc repro locally to keep the tutorial simple. In production, push only the new labeled data with dvc add data/raw and dvc push, then let your CI/CD pipeline retrain the model on the Kubernetes cluster (Part 3 - Serve and deploy).
Scale annotation to the cloud: Deploy Label Studio on a shared server, in a container. The XML labeling interface, export format, and DVC retraining steps stay the same.

Next steps¶

Clean up your resources

Now that you've completed the guide, see the Clean up guide for comprehensive instructions on removing all resources you created:

Local Git repository and DVC cache
Python virtual environment
Cloud storage bucket
Container registry and Docker images
Kubernetes cluster and deployments
CI/CD pipeline configurations
Self-hosted runners
Label Studio installation and data

This is necessary to return to a clean state on your computer, avoid incurring unnecessary costs, and address potential security concerns.

Warning

Unlike previous parts where you could skip cleanup to continue, we strongly recommend completing the full cleanup after finishing Part 5 to avoid ongoing cloud costs (especially Kubernetes clusters) and potential security risks from exposed resources.