bert-finetuned-ner

Este es un moelo afinado sobre el modelo BSC-LT/roberta-base-bne-capitel-ner sobre conll2002 dataset. Se lora un excelente rendimiento porque el modelo original fue preentrenado con textos en español logrando los siguientes resultados:

• Loss: 0.0950
• Precision: 0.8599
• Recall: 0.8773
• F1: 0.8685
• Accuracy: 0.9787

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
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.1045	1.0	521	0.0932	0.8593	0.8704	0.8648	0.9764
0.0343	2.0	1042	0.0870	0.8616	0.8757	0.8686	0.9781
0.019	3.0	1563	0.0950	0.8599	0.8773	0.8685	0.9787

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 2.20.0
• Tokenizers 0.20.0