Modelpublic

ciriatico/dodfminer_lite-ner_bertimbau-extrato_contrato

sourceHugging Facemitupdated 2y agoView on Hugging Face

0likes9downloads

Model Card

e3_lr2e-05

This model is a fine-tuned version of neuralmind/bert-base-portuguese-cased on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.0753
—Precision: 0.9611
—Recall: 0.9778
—F1: 0.9694
—Accuracy: 0.9817

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: 2e-05
—trainbatchsize: 4
—evalbatchsize: 8
—seed: 42
—gradientaccumulationsteps: 2
—totaltrainbatch_size: 8
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 3
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4195	0.2564	50	0.2315	0.8642	0.8460	0.8550	0.9499
0.2396	0.5128	100	0.1778	0.8971	0.8970	0.8970	0.9517
0.1717	0.7692	150	0.1330	0.9033	0.9323	0.9176	0.9639
0.1249	1.0256	200	0.1090	0.9369	0.9554	0.9460	0.9728
0.0929	1.2821	250	0.1066	0.9397	0.9630	0.9512	0.9739
0.0954	1.5385	300	0.0831	0.9498	0.9670	0.9583	0.9788
0.0858	1.7949	350	0.0844	0.9459	0.9727	0.9591	0.9776
0.0715	2.0513	400	0.0868	0.9512	0.9766	0.9637	0.9796
0.056	2.3077	450	0.0789	0.9616	0.9774	0.9695	0.9818
0.0592	2.5641	500	0.0768	0.9614	0.9783	0.9698	0.9817
0.0607	2.8205	550	0.0753	0.9611	0.9778	0.9694	0.9817

Framework versions

—Transformers 4.45.0
—Pytorch 2.4.1+cu121
—Datasets 3.0.0
—Tokenizers 0.20.0