nominal-groups-recognition-bert-base-spanish-wwm-cased

This model is a fine-tuned version of dccuchile/bert-base-spanish-wwm-cased on the jorgeortizfuentes/spanish_nominal_groups_conll2003 dataset. It achieves the following results on the evaluation set:

• Loss: 0.3568
• Ng Precision: 0.7280
• Ng Recall: 0.7767
• Ng F1: 0.7516
• Ng Number: 3198
• Overall Precision: 0.7280
• Overall Recall: 0.7767
• Overall F1: 0.7516
• Overall Accuracy: 0.8992

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: 8
• eval_batch_size: 8
• seed: 13
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Ng Precision	Ng Recall	Ng F1	Ng Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.3955	1.0	228	0.2778	0.7129	0.7492	0.7306	3198	0.7129	0.7492	0.7306	0.8924
0.2186	2.0	456	0.2763	0.7318	0.7711	0.7509	3198	0.7318	0.7711	0.7509	0.8990
0.1586	3.0	684	0.2960	0.7274	0.7733	0.7496	3198	0.7274	0.7733	0.7496	0.8992
0.119	4.0	912	0.3330	0.7283	0.7727	0.7498	3198	0.7283	0.7727	0.7498	0.8982
0.0943	5.0	1140	0.3568	0.7280	0.7767	0.7516	3198	0.7280	0.7767	0.7516	0.8992

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3