🌟 Buying me coffee is a direct way to show support for this project.

distilbert_finetuned_ai4privacy_v2

This model is a fine-tuned version of distilbert-base-uncased on the English Subset of ai4privacy/pii-masking-200k dataset.

Useage

GitHub Implementation: Ai4Privacy

Model description

This model has been finetuned on the World's largest open source privacy dataset.

The purpose of the trained models is to remove personally identifiable information (PII) from text, especially in the context of AI assistants and LLMs.

The example texts have 54 PII classes (types of sensitive data), targeting 229 discussion subjects / use cases split across business, education, psychology and legal fields, and 5 interactions styles (e.g. casual conversation, formal document, emails etc...).

Take a look at the Github implementation for specific reasearch.

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 5

Class wise metrics

It achieves the following results on the evaluation set:

• Loss: 0.0451

• Overall Precision: 0.9438

• Overall Recall: 0.9663

• Overall F1: 0.9549

• Overall Accuracy: 0.9838

• Accountname F1: 0.9946

• Accountnumber F1: 0.9940

• Age F1: 0.9624

• Amount F1: 0.9643

• Bic F1: 0.9929

• Bitcoinaddress F1: 0.9948

• Buildingnumber F1: 0.9845

• City F1: 0.9955

• Companyname F1: 0.9962

• County F1: 0.9877

• Creditcardcvv F1: 0.9643

• Creditcardissuer F1: 0.9953

• Creditcardnumber F1: 0.9793

• Currency F1: 0.7811

• Currencycode F1: 0.8850

• Currencyname F1: 0.2281

• Currencysymbol F1: 0.9562

• Date F1: 0.9061

• Dob F1: 0.7914

• Email F1: 1.0

• Ethereumaddress F1: 1.0

• Eyecolor F1: 0.9837

• Firstname F1: 0.9846

• Gender F1: 0.9971

• Height F1: 0.9910

• Iban F1: 0.9906

• Ip F1: 0.4349

• Ipv4 F1: 0.8126

• Ipv6 F1: 0.7679

• Jobarea F1: 0.9880

• Jobtitle F1: 0.9991

• Jobtype F1: 0.9777

• Lastname F1: 0.9684

• Litecoinaddress F1: 0.9721

• Mac F1: 1.0

• Maskednumber F1: 0.9635

• Middlename F1: 0.9330

• Nearbygpscoordinate F1: 1.0

• Ordinaldirection F1: 0.9910

• Password F1: 1.0

• Phoneimei F1: 0.9918

• Phonenumber F1: 0.9962

• Pin F1: 0.9477

• Prefix F1: 0.9546

• Secondaryaddress F1: 0.9892

• Sex F1: 0.9876

• Ssn F1: 0.9976

• State F1: 0.9893

• Street F1: 0.9873

• Time F1: 0.9889

• Url F1: 1.0

• Useragent F1: 0.9953

• Username F1: 0.9975

• Vehiclevin F1: 1.0

• Vehiclevrm F1: 1.0

• Zipcode F1: 0.9873

Training results

Training Loss	Epoch	Step	Validation Loss	Overall Precision	Overall Recall	Overall F1	Overall Accuracy	Accountname F1	Accountnumber F1	Age F1	Amount F1	Bic F1	Bitcoinaddress F1	Buildingnumber F1	City F1	Companyname F1	County F1	Creditcardcvv F1	Creditcardissuer F1	Creditcardnumber F1	Currency F1	Currencycode F1	Currencyname F1	Currencysymbol F1	Date F1	Dob F1	Email F1	Ethereumaddress F1	Eyecolor F1	Firstname F1	Gender F1	Height F1	Iban F1	Ip F1	Ipv4 F1	Ipv6 F1	Jobarea F1	Jobtitle F1	Jobtype F1	Lastname F1	Litecoinaddress F1	Mac F1	Maskednumber F1	Middlename F1	Nearbygpscoordinate F1	Ordinaldirection F1	Password F1	Phoneimei F1	Phonenumber F1	Pin F1	Prefix F1	Secondaryaddress F1	Sex F1	Ssn F1	State F1	Street F1	Time F1	Url F1	Useragent F1	Username F1	Vehiclevin F1	Vehiclevrm F1	Zipcode F1
0.6445	1.0	1088	0.3322	0.6449	0.7003	0.6714	0.8900	0.7607	0.8733	0.6576	0.1766	0.25	0.6783	0.3621	0.6005	0.6909	0.5586	0.0	0.2449	0.7095	0.2889	0.0	0.0	0.3902	0.7720	0.0	0.9862	0.8011	0.5088	0.7740	0.7118	0.5434	0.8088	0.0	0.8303	0.7562	0.5318	0.7294	0.4681	0.6779	0.0	0.8909	0.0	0.0107	0.9985	0.4000	0.7307	0.9057	0.8618	0.0	0.9127	0.8235	0.9211	0.8026	0.4656	0.6390	0.9383	0.9775	0.8868	0.8201	0.4526	0.0550	0.5368
0.222	2.0	2176	0.1259	0.8170	0.8747	0.8449	0.9478	0.9708	0.9813	0.7638	0.7427	0.7837	0.8908	0.8833	0.8747	0.9814	0.8749	0.7601	0.9777	0.8834	0.5372	0.4828	0.0056	0.7785	0.8149	0.3140	0.9956	0.9935	0.9101	0.9270	0.9450	0.9853	0.9253	0.0650	0.0084	0.7962	0.9013	0.9446	0.9203	0.8555	0.6885	1.0	0.7152	0.6442	1.0	0.9623	0.9349	0.9905	0.9782	0.7656	0.9324	0.9903	0.9736	0.9274	0.8520	0.9138	0.9678	0.9922	0.9893	0.9804	0.9646	0.8556	0.8385
0.1331	3.0	3264	0.0773	0.9133	0.9371	0.9250	0.9654	0.9822	0.9815	0.9196	0.8852	0.9718	0.9785	0.9215	0.9757	0.9935	0.9651	0.8742	0.9921	0.9438	0.7568	0.7710	0.0	0.8998	0.7895	0.6578	0.9994	1.0	0.9554	0.9525	0.9823	0.9910	0.9866	0.0435	0.8293	0.7824	0.9671	0.9794	0.9571	0.9447	0.9141	1.0	0.8825	0.7988	1.0	0.9797	0.9921	0.9932	0.9943	0.8726	0.9401	0.9860	0.9792	0.9928	0.9740	0.9604	0.9730	0.9983	0.9964	0.9959	0.9890	0.9774	0.9247
0.0847	4.0	4352	0.0503	0.9368	0.9614	0.9489	0.9789	0.9955	0.9949	0.9573	0.9480	0.9929	0.9846	0.9808	0.9927	0.9962	0.9811	0.9436	0.9953	0.9695	0.7826	0.8713	0.1653	0.9458	0.8782	0.7996	1.0	1.0	0.9809	0.9816	0.9941	0.9910	0.9906	0.3389	0.8364	0.7066	0.9862	1.0	0.9795	0.9637	0.9429	1.0	0.9438	0.9165	1.0	0.9864	1.0	0.9932	0.9962	0.9352	0.9483	0.9860	0.9866	0.9976	0.9884	0.9827	0.9881	1.0	0.9953	0.9975	0.9945	0.9915	0.9841
0.0557	5.0	5440	0.0451	0.9438	0.9663	0.9549	0.9838	0.9946	0.9940	0.9624	0.9643	0.9929	0.9948	0.9845	0.9955	0.9962	0.9877	0.9643	0.9953	0.9793	0.7811	0.8850	0.2281	0.9562	0.9061	0.7914	1.0	1.0	0.9837	0.9846	0.9971	0.9910	0.9906	0.4349	0.8126	0.7679	0.9880	0.9991	0.9777	0.9684	0.9721	1.0	0.9635	0.9330	1.0	0.9910	1.0	0.9918	0.9962	0.9477	0.9546	0.9892	0.9876	0.9976	0.9893	0.9873	0.9889	1.0	0.9953	0.9975	1.0	1.0	0.9873

Framework versions

• Transformers 4.35.0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.14.1