hucruz/pautas

custom-ner-model

This model is a fine-tuned version of dccuchile/distilbert-base-spanish-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2706
• Precision: 0.8139
• Recall: 0.8624
• F1: 0.8374
• Accuracy: 0.9376

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	55	0.2457	0.7716	0.8211	0.7956	0.9346
No log	2.0	110	0.3008	0.75	0.8257	0.7860	0.9152
No log	3.0	165	0.2658	0.7712	0.8349	0.8018	0.9322
No log	4.0	220	0.2580	0.7583	0.8349	0.7948	0.9322
No log	5.0	275	0.2610	0.8079	0.8486	0.8277	0.9346
No log	6.0	330	0.2660	0.7699	0.8440	0.8053	0.9303
No log	7.0	385	0.2910	0.7830	0.8440	0.8124	0.9261
No log	8.0	440	0.2577	0.7913	0.8349	0.8125	0.9376
No log	9.0	495	0.2629	0.7948	0.8349	0.8143	0.9346
0.0437	10.0	550	0.2745	0.8070	0.8440	0.8251	0.9328
0.0437	11.0	605	0.2733	0.7860	0.8257	0.8054	0.9316
0.0437	12.0	660	0.2641	0.7965	0.8440	0.8196	0.9310
0.0437	13.0	715	0.2716	0.8139	0.8624	0.8374	0.9334
0.0437	14.0	770	0.2765	0.8210	0.8624	0.8412	0.9334
0.0437	15.0	825	0.2775	0.8253	0.8670	0.8456	0.9346
0.0437	16.0	880	0.2737	0.7897	0.8440	0.8160	0.9340
0.0437	17.0	935	0.2691	0.8043	0.8486	0.8259	0.9388
0.0437	18.0	990	0.2694	0.8246	0.8624	0.8430	0.9382
0.0157	19.0	1045	0.2732	0.8069	0.8624	0.8337	0.9346
0.0157	20.0	1100	0.2706	0.8139	0.8624	0.8374	0.9376

Framework versions

• Transformers 4.26.0
• Pytorch 1.13.1+cu116
• Datasets 2.9.0
• Tokenizers 0.13.2