maskara-tiny

This model is a fine-tuned version of somukandula/maskara-tiny on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5828
• Precision: 0.8369
• Recall: 0.9271
• F1: 0.8797
• Overall F1: 0.8172
• Address F1: 0.6667
• Api Key F1: 1.0
• Credit Card F1: 0.0
• Date Of Birth F1: 1.0
• Driver License F1: 0.9875
• Email F1: 1.0
• Ip Address F1: 1.0
• Password F1: 0.0
• Person Name F1: 0.8378
• Phone F1: 0.8267
• Ssn F1: 0.0
• Username F1: 1.0
• Aadhaar F1: 0.9637
• Pan Card F1: 1.0
• Passport F1: 0.0
• Upi Id F1: 1.0
• Vehicle Reg F1: 0.975
• Indian Id F1: 0.9847

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Overall F1	Address F1	Api Key F1	Credit Card F1	Date Of Birth F1	Driver License F1	Email F1	Ip Address F1	Password F1	Person Name F1	Phone F1	Ssn F1	Username F1	Aadhaar F1	Pan Card F1	Passport F1	Upi Id F1	Vehicle Reg F1	Indian Id F1
0.0006	1.0	14283	0.6201	0.7070	0.7883	0.7454	0.7770	0.9996	0.9888	0.2740	0.9882	0.5042	0.6678	0.9995	0.5025	0.8300	0.6457	1.0	0.9623	0.6661	0.8804	0.9544	0.9077	0.4370	0.7228
0.0003	2.0	28566	0.4654	0.7210	0.8069	0.7615	0.7900	0.9998	0.9880	0.3006	0.9835	0.5868	0.7152	0.9667	0.6032	0.8403	0.6424	1.0	0.9833	0.6575	0.8424	0.8378	0.9728	0.5097	0.7456
0.0003	3.0	42849	0.5373	0.7026	0.8087	0.7519	0.7755	0.9992	0.9857	0.2915	0.9859	0.5949	0.7581	0.9446	0.4776	0.8277	0.6466	1.0	0.9880	0.6732	0.8555	0.8258	0.9161	0.4128	0.7144

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1