distilbert-base-uncased-pii-finance

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

• Loss: 0.0652
• Overall Precision: 0.7981
• Overall Recall: 0.8151
• Overall F1: 0.8065
• Overall Accuracy: 0.9811
• Bic F1: 0.8031
• Companyname F1: 0.7576
• Creditcardcvv F1: 0.5929
• Creditcardnumber F1: 0.8070
• Date F1: 0.8121
• Dob F1: 0.8606
• Email F1: 0.8827
• Firstname F1: 0.3405
• Iban F1: 0.7970
• Ipv4 F1: 0.9398
• Ipv6 F1: 0.8089
• Lastname F1: 0.2717
• Nearbygpscoordinate F1: 0.4321
• Password F1: 0.5250
• Phonenumber F1: 0.8580
• Pin F1: 0.7881
• Ssn F1: 0.8408
• Street F1: 0.8066
• Time F1: 0.6761
• Username F1: 0.8691

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
• lr_scheduler_warmup_ratio: 0.2
• num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Overall Precision	Overall Recall	Overall F1	Overall Accuracy	Bic F1	Companyname F1	Creditcardcvv F1	Creditcardnumber F1	Date F1	Dob F1	Email F1	Firstname F1	Iban F1	Ipv4 F1	Ipv6 F1	Lastname F1	Nearbygpscoordinate F1	Password F1	Phonenumber F1	Pin F1	Ssn F1	Street F1	Time F1	Username F1
0.8108	1.0	563	0.1036	0.6987	0.7593	0.7277	0.9726	0.1895	0.6896	0.0	0.0352	0.7602	0.7419	0.8776	0.0	0.5259	0.6442	0.4383	0.0	0.3790	0.04	0.7697	0.0	0.0435	0.7134	0.4934	0.8312
0.0888	2.0	1126	0.0736	0.7366	0.8299	0.7805	0.9780	0.8304	0.7459	0.1584	0.6792	0.7845	0.8224	0.8846	0.0	0.7454	0.9440	0.7467	0.0498	0.2077	0.4024	0.8351	0.6197	0.7761	0.8020	0.6392	0.8356
0.0649	3.0	1689	0.0665	0.7602	0.8371	0.7968	0.9797	0.8656	0.7445	0.5714	0.6644	0.8027	0.8320	0.8793	0.2841	0.8317	0.9291	0.7054	0.2823	0.6	0.5747	0.8537	0.7336	0.7491	0.8053	0.7002	0.8590
0.051	4.0	2252	0.0652	0.7981	0.8151	0.8065	0.9811	0.8031	0.7576	0.5929	0.8070	0.8121	0.8606	0.8827	0.3405	0.7970	0.9398	0.8089	0.2717	0.4321	0.5250	0.8580	0.7881	0.8408	0.8066	0.6761	0.8691
0.0415	5.0	2815	0.0664	0.7805	0.8488	0.8132	0.9812	0.8606	0.7735	0.6353	0.7584	0.8225	0.8257	0.8626	0.4258	0.8687	0.9298	0.8241	0.3284	0.5104	0.6000	0.8607	0.7444	0.8346	0.8227	0.7345	0.8604
0.0353	6.0	3378	0.0667	0.7954	0.8403	0.8172	0.9816	0.8828	0.7780	0.6275	0.8033	0.8255	0.8533	0.8510	0.4687	0.8256	0.9327	0.8128	0.3348	0.5638	0.5618	0.8574	0.7549	0.84	0.8142	0.7445	0.8704
0.0303	7.0	3941	0.0696	0.7939	0.8475	0.8198	0.9816	0.8845	0.7844	0.6349	0.7692	0.8269	0.8506	0.8644	0.4519	0.8425	0.9257	0.8333	0.3543	0.5654	0.5618	0.8636	0.7605	0.8537	0.8236	0.7440	0.8648

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1