Gliner PII ITA

🚀 Gliner - General Model for PII Identification

This model is designed to identify and classify Personally Identifiable Information (PII) within text documents. It is trained to be general, capable of recognizing a wide range of PII tags, with a specialization in identifying the most common and relevant tags in administrative, legal, and financial fields.

✨ Features

Application Areas

It is particularly effective in analyzing legal documents, contracts, financial records, and administrative documents. Correctly identifying and classifying personal information is crucial for compliance with privacy regulations (such as the GDPR).

Supported Tags

The model has been trained with a total of 11706 different tags. Some examples of tags the model can recognize include:

• Customer Name: Identifies the full name of a customer.
• Place of Birth: Recognizes the municipality where a person was born.
• Residential Address: Recognizes a person's residential address.
• Tax Code: Recognizes the tax code of an individual or a company.
• Document Number: Identifies identification document numbers such as passports, identity cards, etc.
• Transaction Amount: Recognizes monetary amounts related to financial transactions.
• Cadastral Parcel Number: Identifies the cadastral parcel number related to a property.
• Company Name: Recognizes the official name of a company.
• IBAN: Identifies bank account numbers in IBAN format.
• IP Address: Recognizes IP addresses associated with users or devices.

Performance

The model has been trained on a large dataset containing approximately 200K manually annotated examples for each type of PII.

Limitations

• Data Variability: Gliner may have difficulty identifying PII in unstructured documents or those with formats very different from the training set.
• Rare Tags: The model may show reduced performance on less common tags or those not well - represented in the training set.
• Bias: Like all machine learning models, Gliner may be subject to bias from the training dataset.

📦 Installation

To use this model, you must install the GLiNER Python library:

!pip install gliner

💻 Usage Examples

Basic Usage

from gliner import GLiNER

model = GLiNER.from_pretrained("DeepMount00/GLiNER_PII_ITA")

text = """..."""

labels = ["label1", "label2"]

entities = model.predict_entities(text, labels)

for entity in entities:
    print(entity["text"], "=>", entity["label"])

📄 License

This project is licensed under the Apache - 2.0 license.