DeBERTa Fine-tuned for PII Detection (ONNX)

This is an ONNX-exported version of a DeBERTa model fine-tuned for PII (Personally Identifiable Information) token classification. The model is based on lakshyakh93/deberta_finetuned_pii and has been converted to ONNX format for optimized inference.

Model Description

Model Type: DeBERTa
Task: Token Classification (Named Entity Recognition)
Framework: ONNX
Original Model: lakshyakh93/deberta_finetuned_pii

PII Types Detected

This model can identify a wide range of PII elements, including but not limited to:

Personal Names (First, Middle, Last names)
Contact Information (Email, Phone Numbers)
Addresses (Street, City, State, ZIP)
Financial Information (Credit Card, Account Numbers, IBAN)
Digital Identifiers (IP Addresses, URLs, Usernames)
Cryptocurrency Addresses (Bitcoin, Ethereum, Litecoin)
Identification Numbers (SSN, VIN, IMEI)
Professional Information (Job Titles, Company Names)
Temporal Information (Dates, Times)
And many more

Model Architecture

Hidden Size: 768
Number of Hidden Layers: 12
Number of Attention Heads: 12
Maximum Sequence Length: 512

Usage

This ONNX model can be used with any ONNX-compatible runtime for efficient inference in production environments.

Export Details

The model has been exported to ONNX format while preserving the original model's capabilities for token classification tasks.

License

Please refer to the original model's license terms.