Update README.md
Browse files
README.md
CHANGED
|
@@ -6,23 +6,22 @@ license: mit
|
|
| 6 |
datasets:
|
| 7 |
- ACE2005
|
| 8 |
---
|
| 9 |
-
# Arabic NER
|
| 10 |
-
|
| 11 |
-
### Model
|
| 12 |
- NER BIO tagging model based on [GigaBERTv4](https://huggingface.co/lanwuwei/GigaBERT-v4-Arabic-and-English).
|
| 13 |
- ACE2005 Training data: English + Arabic
|
| 14 |
- [NER tags](https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/english-entities-guidelines-v6.6.pdf) including: PER, VEH, GPE, WEA, ORG, LOC, FAC
|
| 15 |
|
| 16 |
-
##
|
| 17 |
- learning_rate=2e-5
|
| 18 |
- num_train_epochs=10
|
| 19 |
- weight_decay=0.01
|
| 20 |
-
|
|
|
|
| 21 |
| Language | Arabic | English |
|
| 22 |
|:----:|:-----------:|:----:|
|
| 23 |
| | 89.4 | 88.8 |
|
| 24 |
|
| 25 |
-
##
|
| 26 |
```python
|
| 27 |
>>> from transformers import pipeline, AutoModelForTokenClassification, AutoTokenizer
|
| 28 |
|
|
|
|
| 6 |
datasets:
|
| 7 |
- ACE2005
|
| 8 |
---
|
| 9 |
+
# Arabic NER Model
|
|
|
|
|
|
|
| 10 |
- NER BIO tagging model based on [GigaBERTv4](https://huggingface.co/lanwuwei/GigaBERT-v4-Arabic-and-English).
|
| 11 |
- ACE2005 Training data: English + Arabic
|
| 12 |
- [NER tags](https://www.ldc.upenn.edu/sites/www.ldc.upenn.edu/files/english-entities-guidelines-v6.6.pdf) including: PER, VEH, GPE, WEA, ORG, LOC, FAC
|
| 13 |
|
| 14 |
+
## Hyperparameters
|
| 15 |
- learning_rate=2e-5
|
| 16 |
- num_train_epochs=10
|
| 17 |
- weight_decay=0.01
|
| 18 |
+
|
| 19 |
+
## ACE2005 Evaluation results
|
| 20 |
| Language | Arabic | English |
|
| 21 |
|:----:|:-----------:|:----:|
|
| 22 |
| | 89.4 | 88.8 |
|
| 23 |
|
| 24 |
+
##How to use
|
| 25 |
```python
|
| 26 |
>>> from transformers import pipeline, AutoModelForTokenClassification, AutoTokenizer
|
| 27 |
|