PII-Detection

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: 0.0643
• Overall Precision: 0.9370
• Overall Recall: 0.9502
• Overall F1: 0.9436
• Overall Accuracy: 0.9823
• Accountname F1: 0.9835
• Accountnumber F1: 0.9792
• Amount F1: 0.9439
• Bic F1: 0.8741
• Bitcoinaddress F1: 0.9478
• Buildingnumber F1: 0.6465
• City F1: 0.9865
• Company Name F1: 0.9583
• County F1: 0.9811
• Creditcardcvv F1: 0.9007
• Creditcardissuer F1: 0.9302
• Creditcardnumber F1: 0.8662
• Currency F1: 0.7101
• Currencycode F1: 0.7445
• Currencyname F1: 0.4375
• Currencysymbol F1: 0.5600
• Date F1: 0.9873
• Displayname F1: 0.4706
• Email F1: 0.9991
• Ethereumaddress F1: 0.9808
• Firstname F1: 0.8912
• Fullname F1: 0.9867
• Gender F1: 0.8462
• Iban F1: 0.9947
• Ip F1: 0.2642
• Ipv4 F1: 0.8165
• Ipv6 F1: 0.6330
• Jobarea F1: 0.9734
• Jobdescriptor F1: 0.8723
• Jobtitle F1: 0.9760
• Jobtype F1: 0.9150
• Lastname F1: 0.7589
• Litecoinaddress F1: 0.9123
• Mac F1: 1.0
• Maskednumber F1: 0.8000
• Middlename F1: 0.8033
• Name F1: 0.9966
• Nearbygpscoordinate F1: 1.0
• Number F1: 0.9159
• Ordinaldirection F1: 0.0
• Password F1: 0.9497
• Phoneimei F1: 0.9756
• Phone Number F1: 0.9244
• Pin F1: 0.8889
• Prefix F1: 0.9009
• Secondaryaddress F1: 0.9939
• Sex F1: 0.9032
• Sextype F1: 0.0
• Ssn F1: 0.8992
• State F1: 0.9931
• Street F1: 0.6906
• Streetaddress F1: 0.8523
• Suffix F1: 0.9026
• Time F1: 0.9796
• Url F1: 0.9973
• Useragent F1: 0.9839
• Username F1: 0.8900
• Vehiclevin F1: 0.9612
• Vehiclevrm F1: 0.9697
• Zipcode F1: 0.9387

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	Amount F1	Bic F1	Bitcoinaddress F1	Buildingnumber F1	City F1	Company Name F1	County F1	Creditcardcvv F1	Creditcardissuer F1	Creditcardnumber F1	Currency F1	Currencycode F1	Currencyname F1	Currencysymbol F1	Date F1	Displayname F1	Email F1	Ethereumaddress F1	Firstname F1	Fullname F1	Gender F1	Iban F1	Ip F1	Ipv4 F1	Ipv6 F1	Jobarea F1	Jobdescriptor F1	Jobtitle F1	Jobtype F1	Lastname F1	Litecoinaddress F1	Mac F1	Maskednumber F1	Middlename F1	Name F1	Nearbygpscoordinate F1	Number F1	Ordinaldirection F1	Password F1	Phoneimei F1	Phone Number F1	Pin F1	Prefix F1	Secondaryaddress F1	Sex F1	Sextype F1	Ssn F1	State F1	Street F1	Streetaddress F1	Suffix F1	Time F1	Url F1	Useragent F1	Username F1	Vehiclevin F1	Vehiclevrm F1	Zipcode F1
0.1445	1.0	1337	0.0873	0.8766	0.9038	0.8900	0.9710	0.9132	0.9245	0.8316	0.6260	0.7211	0.5392	0.9670	0.8041	0.9658	0.6462	0.7333	0.6833	0.5885	0.4341	0.016	0.0513	0.9726	0.0	0.9986	0.8778	0.7974	0.9812	0.5905	0.8744	0.0405	0.7125	0.7616	0.9148	0.0	0.7961	0.5644	0.5821	0.6442	0.9573	0.2375	0.1486	0.9865	0.3077	0.7009	0.0	0.8883	0.9697	0.7985	0.6387	0.8256	0.8772	0.8742	0.0	0.8593	0.9870	0.5492	0.8496	0.4966	0.9632	0.9754	0.9370	0.8839	0.8857	0.8	0.8691
0.0599	2.0	2674	0.0573	0.9218	0.9369	0.9293	0.9786	0.9669	0.9679	0.9179	0.8429	0.78	0.4982	0.9831	0.9375	0.9774	0.8873	0.9153	0.8418	0.7326	0.6621	0.1102	0.5634	0.975	0.0	0.9972	0.9202	0.8603	0.9859	0.7925	0.9684	0.1164	0.7847	0.7129	0.9626	0.6752	0.9445	0.8734	0.6809	0.6931	0.9811	0.7468	0.6905	0.9943	1.0	0.8165	0.0	0.9385	0.9756	0.8465	0.8190	0.8899	0.9878	0.8861	0.0	0.8769	0.9905	0.6496	0.8784	0.8513	0.9662	0.9932	0.9490	0.9149	0.9394	0.9265	0.8732
0.0411	3.0	4011	0.0517	0.9146	0.9444	0.9292	0.9804	0.9703	0.9837	0.9269	0.8939	0.8806	0.615	0.9837	0.9220	0.9811	0.8742	0.9249	0.8497	0.6351	0.6970	0.4037	0.5195	0.9824	0.3956	0.9989	0.9761	0.8779	0.9851	0.8081	0.9894	0.1043	0.8622	0.5398	0.9655	0.8333	0.9742	0.92	0.7368	0.8432	0.9905	0.7389	0.7661	0.9947	1.0	0.8505	0.0	0.9435	1.0	0.9204	0.8491	0.8920	0.9878	0.8889	0.0	0.9538	0.9896	0.6269	0.7057	0.8900	0.9796	0.9945	0.9641	0.8623	0.9606	0.9143	0.9672
0.0226	4.0	5348	0.0599	0.9349	0.9482	0.9415	0.9824	0.9835	0.9746	0.9041	0.8939	0.8963	0.6707	0.9871	0.9141	0.9848	0.8707	0.9302	0.8702	0.6916	0.6618	0.3830	0.5195	0.9848	0.4651	0.9986	0.9581	0.8872	0.9862	0.8776	0.9892	0.2561	0.8682	0.7172	0.9726	0.8763	0.9814	0.9322	0.7266	0.7826	0.9952	0.8193	0.7256	0.9960	1.0	0.9014	0.0	0.9494	0.9816	0.9238	0.8462	0.9002	0.9939	0.8987	0.0	0.9091	0.9922	0.6997	0.8472	0.8821	0.9796	0.9973	0.9758	0.8940	0.9538	0.9624	0.9485
0.0129	5.0	6685	0.0643	0.9370	0.9502	0.9436	0.9823	0.9835	0.9792	0.9439	0.8741	0.9478	0.6465	0.9865	0.9583	0.9811	0.9007	0.9302	0.8662	0.7101	0.7445	0.4375	0.5600	0.9873	0.4706	0.9991	0.9808	0.8912	0.9867	0.8462	0.9947	0.2642	0.8165	0.6330	0.9734	0.8723	0.9760	0.9150	0.7589	0.9123	1.0	0.8000	0.8033	0.9966	1.0	0.9159	0.0	0.9497	0.9756	0.9244	0.8889	0.9009	0.9939	0.9032	0.0	0.8992	0.9931	0.6906	0.8523	0.9026	0.9796	0.9973	0.9839	0.8900	0.9612	0.9697	0.9387

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1