eriktks/conll2002
Updated • 970 • 11
How to use ifis/BETO-finetuned-ner-3 with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("token-classification", model="ifis/BETO-finetuned-ner-3") # Load model directly
from transformers import AutoTokenizer, AutoModelForTokenClassification
tokenizer = AutoTokenizer.from_pretrained("ifis/BETO-finetuned-ner-3")
model = AutoModelForTokenClassification.from_pretrained("ifis/BETO-finetuned-ner-3")Este es un modelo fine-tuned de NazaGara/NER-fine-tuned-BETO con el dataset conll2002. Los resultados fueron los siguientes:
Es un modelo de secuencias basado en BETO para la tarea de etiquetado de secuencias (NER - Named Entity Recognition) y ajustado (fine-tuned). Utiliza transformadores para identificar y clasificar entidades nombradas en el texto.
The following hyperparameters were used during training:
[13880 / 13880, Epoch 20/20]
| Step | Training Loss | Validation Loss |
|---|---|---|
| 250 | No log | 0.155875 |
| 500 | 0.063300 | 0.149275 |
| 750 | 0.063300 | 0.149796 |
| 1000 | 0.041700 | 0.157160 |
| 1250 | 0.041700 | 0.159034 |
| 1500 | 0.032700 | 0.178371 |
| 1750 | 0.032700 | 0.188239 |
| 2000 | 0.024100 | 0.174550 |
| 2250 | 0.024100 | 0.177991 |
| 2500 | 0.021200 | 0.196993 |
| 2750 | 0.017200 | 0.171864 |
| 3000 | 0.014200 | 0.221017 |
| 3250 | 0.014200 | 0.188359 |
| 3500 | 0.013500 | 0.202872 |
| 3750 | 0.012900 | 0.216551 |
| 4000 | 0.012300 | 0.210710 |
| 4250 | 0.010100 | 0.226150 |
| 4500 | 0.007200 | 0.211124 |
| 4750 | 0.007200 | 0.233601 |
| 5000 | 0.006400 | 0.240447 |
| 5250 | 0.006400 | 0.218564 |
| 5500 | 0.006200 | 0.220331 |
| 5750 | 0.006200 | 0.237083 |
| 6000 | 0.004700 | 0.243056 |
| 6250 | 0.004700 | 0.240422 |
| 6500 | 0.004700 | 0.250421 |
| 6750 | 0.004500 | 0.258444 |
| 7000 | 0.004500 | 0.245971 |
| 7250 | 0.004500 | 0.256681 |
| 7500 | 0.003500 | 0.246265 |
| 7750 | 0.003500 | 0.256869 |
| 8000 | 0.002900 | 0.255373 |
| 8250 | 0.002900 | 0.251113 |
| 8500 | 0.002800 | 0.264475 |
| 8750 | 0.002800 | 0.260816 |
| 9000 | 0.003100 | 0.285076 |
| 9250 | 0.003100 | 0.275611 |
| 9500 | 0.002700 | 0.284239 |
| 9750 | 0.002700 | 0.285485 |
| 10000 | 0.002400 | 0.292293 |
| 10250 | 0.002400 | 0.274808 |
| 10500 | 0.002000 | 0.288339 |
| 10750 | 0.002000 | 0.285888 |
| 11000 | 0.001700 | 0.296517 |
| 11250 | 0.001700 | 0.284072 |
| 11500 | 0.001800 | 0.285594 |
| 11750 | 0.001800 | 0.283680 |
| 12000 | 0.001900 | 0.282890 |
| 12250 | 0.001900 | 0.279963 |
| 12500 | 0.001700 | 0.280620 |
| 12750 | 0.001700 | 0.280996 |
| 13000 | 0.001400 | 0.279926 |
| 13250 | 0.001400 | 0.282571 |
| 13500 | 0.001300 | 0.282674 |
| 13750 | 0.001300 | 0.282024 |