Modelpublic

PLN-T4-J-D-W/beto-finetuned-ner-1

sourceHugging Faceupdated 2y agoView on Hugging Face

0likes8downloads

Model Card

beto-finetuned-ner-1

Este es modelo resultado de un finetuning de NazaGara/NER-fine-tuned-BETO sobre el conll2002 dataset. Los siguientes son los resultados sobre el conjunto de evaluación:

—Loss: 0.002421
—Precision: 0.861199
—Recall: 0.871094
—F1: 0.8851
—Accuracy: 0,972756

Model description

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
—weight_decay: 0.001
—num_epochs: 8

Training results

Epoch	Training Loss	Validation Loss	Precision	Recall	F1	Accuracy
1	0.004500	0.271499	0.854365	0.868107	0.861181	0.971268
2	0.004000	0.283811	0.839605	0.840763	0.840184	0.966170
3	0.003900	0.261076	0.849651	0.867417	0.858442	0.970664
4	0.002600	0.277270	0.858379	0.866268	0.862306	0.971702
5	0.002000	0.270548	0.859149	0.871783	0.865420	0.971563
6	0.001800	0.279797	0.857305	0.868336	0.862785	0.971609
7	0.001800	0.281091	0.857467	0.868107	0.862754	0.971966
8	0.001100	0.284128	0.861199	0.871094	0.866118	0.972756