PrERT-CNM v4 PrivacyBERT is a Transformers sequence-classification model prepared from the local checkpoint at artifacts\phase-3-privacybert\classifier_checkpoint\privacybert.
Use this model for text classification in the privacy/CNM workflow it was trained for. It is intended for research and application prototyping unless your own validation shows it is suitable for production use.
1from transformers import pipeline
2
3classifier = pipeline("text-classification", model="JayNightmare/PrERT-CNM-v4-privacybert", top_k=None)
4scores = classifier("Paste text to classify.")
5print(scores)
Add the final held-out metrics before publishing if they are available. Include dataset split details, label distribution, and any thresholding used by downstream consumers.
The model can be sensitive to domain shift, ambiguous language, long inputs, and label definitions that differ from the training data. Review outputs before using them in automated decisions.
