PII-Detection-V2.1

deepaksiloka/PII-Detection-V2.1

PII-Detection-V2.1 is a token classification model from deepaksiloka. Use it when you need labels on individual words, such as names. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

0% of all-time downloads

All-time downloads

16.4K

Public

Parameters

66.4M

2.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors266 MB · 100%

At a glance

Task: Token Classification
Library: transformers
License: apache-2.0
Model type: distilbert
Access: Public
Created: Oct 11, 2024
Updated: Oct 14, 2024
SHA: cff95d07

Try a prompt

Base models

Model card

From the Hugging Face model README

PII-Detection-V2.1

This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0331
Overall Precision: 0.9482
Overall Recall: 0.9574
Overall F1: 0.9528
Overall Accuracy: 0.9926
Accountname F1: 0.9939
Accountnumber F1: 0.9879
Buildingnumber F1: 0.8059
City F1: 0.9729
Companyname F1: 0.9773
County F1: 0.9463
Creditcardcvv F1: 0.8970
Creditcardissuer F1: 0.9565
Creditcardnumber F1: 0.8770
Email F1: 0.9981
Firstname F1: 0.9324
Fullname F1: 0.9851
Iban F1: 0.9834
Lastname F1: 0.8744
Middlename F1: 0.8390
Name F1: 0.9972
Number F1: 0.9684
Phonenumber F1: 0.9788
Pin F1: 0.9017
Secondaryaddress F1: 0.9892
State F1: 0.9421
Street F1: 0.8617
Streetaddress F1: 0.7533
Url F1: 0.9977
Username F1: 0.9654

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Overall Precision	Overall Recall	Overall F1	Overall Accuracy	Accountname F1	Accountnumber F1	Buildingnumber F1	City F1	Companyname F1	County F1	Creditcardcvv F1	Creditcardissuer F1	Creditcardnumber F1	Email F1	Firstname F1	Fullname F1	Iban F1	Lastname F1	Middlename F1	Name F1	Number F1	Phonenumber F1	Pin F1	Secondaryaddress F1	State F1	Street F1	Streetaddress F1	Url F1	Username F1
0.0445	1.0	2031	0.0375	0.8924	0.9268	0.9093	0.9885	0.9781	0.9675	0.7260	0.9391	0.9347	0.8747	0.8221	0.8962	0.8239	0.9948	0.8764	0.9804	0.9528	0.7683	0.6548	0.9889	0.8223	0.9277	0.7938	0.9853	0.8633	0.7646	0.4597	0.9937	0.9427
0.0266	2.0	4062	0.0296	0.9245	0.9455	0.9349	0.9908	0.9900	0.9810	0.7546	0.9639	0.9574	0.9085	0.8370	0.9375	0.8809	0.9979	0.9094	0.9824	0.9785	0.8299	0.8111	0.9938	0.9247	0.9523	0.8640	0.9826	0.9163	0.7605	0.6372	0.9977	0.9599
0.0148	3.0	6093	0.0277	0.9414	0.9529	0.9471	0.9921	0.9948	0.9863	0.7876	0.9689	0.9624	0.9324	0.8883	0.9537	0.8795	0.9979	0.9252	0.9849	0.9840	0.8515	0.8310	0.9946	0.9506	0.9675	0.8685	0.9875	0.9325	0.8355	0.7560	0.9973	0.9685
0.0095	4.0	8124	0.0301	0.9438	0.9536	0.9487	0.9921	0.9913	0.9859	0.8018	0.9742	0.9652	0.9443	0.8982	0.9508	0.8784	0.9986	0.9281	0.9842	0.9828	0.8584	0.8294	0.9952	0.9681	0.9629	0.8889	0.9875	0.9374	0.8430	0.7522	0.9980	0.9457
0.0038	5.0	10155	0.0331	0.9482	0.9574	0.9528	0.9926	0.9939	0.9879	0.8059	0.9729	0.9773	0.9463	0.8970	0.9565	0.8770	0.9981	0.9324	0.9851	0.9834	0.8744	0.8390	0.9972	0.9684	0.9788	0.9017	0.9892	0.9421	0.8617	0.7533	0.9977	0.9654

Framework versions

Transformers 4.45.2
Pytorch 2.2.0
Datasets 3.0.1
Tokenizers 0.20.1

Task: Token Classification
Library: transformers
Type: distilbert
License: apache-2.0
Created: Oct 11, 2024
Updated: Oct 14, 2024