Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# XLM-RoBERTa large for QA on Vietnamese languages (also support various languages)
|
| 2 |
+
|
| 3 |
+
## Overview
|
| 4 |
+
|
| 5 |
+
- Language model: xlm-roberta-large
|
| 6 |
+
- Fine-tune: [deepset/xlm-roberta-large-squad2](https://huggingface.co/deepset/xlm-roberta-large-squad2)
|
| 7 |
+
- Language: Vietnamese
|
| 8 |
+
- Downstream-task: Extractive QA
|
| 9 |
+
- Dataset: [mailong25/bert-vietnamese-question-answering](https://github.com/mailong25/bert-vietnamese-question-answering/tree/master/dataset)
|
| 10 |
+
- Training data: train-v2.0.json (SQuAD 2.0 format)
|
| 11 |
+
- Eval data: dev-v2.0.json (SQuAD 2.0 format)
|
| 12 |
+
- Infrastructure: 1x Tesla P100 (Google Colab)
|
| 13 |
+
|
| 14 |
+
## Performance
|
| 15 |
+
|
| 16 |
+
Evaluated on dev-v2.0.json
|
| 17 |
+
```
|
| 18 |
+
exact: 136 / 141
|
| 19 |
+
f1: 0.9692671394799054
|
| 20 |
+
```
|
| 21 |
+
|
| 22 |
+
## Author
|
| 23 |
+
|
| 24 |
+
An Pham: ancs21.ps [at] gmail.com
|
| 25 |
+
|
| 26 |
+
## License
|
| 27 |
+
|
| 28 |
+
MIT
|