|
|
--- |
|
|
language: en |
|
|
license: mit |
|
|
library_name: transformers |
|
|
pipeline_tag: question-answering |
|
|
tags: |
|
|
- question-answering |
|
|
- roberta |
|
|
- squad2 |
|
|
- deepset |
|
|
dataset: |
|
|
- squad2 |
|
|
--- |
|
|
|
|
|
# CTION-QA: Question Answering Model |
|
|
|
|
|
## Model Description |
|
|
A Question Answering (Q&A) model is a transformer-based NLP model trained to understand a given context and accurately extract or generate answers to user questions from that text. |
|
|
It is fine-tuned on the **SQuAD 2.0** dataset for extractive question answering. |
|
|
|
|
|
## Performance |
|
|
| Metric | Score | |
|
|
|------|------| |
|
|
| Exact Match | 76.9 | |
|
|
| F1 Score | 79.8 | |
|
|
| Context Length | 512 | |
|
|
|
|
|
## Usage |
|
|
```python |
|
|
from transformers import pipeline |
|
|
|
|
|
qa = pipeline( |
|
|
"question-answering", |
|
|
model="hariprabhakaran45/CTION-QA" |
|
|
) |
|
|
|
|
|
result = qa( |
|
|
question="Who is the Eiffel Tower named after?", |
|
|
context="The Eiffel Tower is named after Gustave Eiffel." |
|
|
) |
|
|
|
|
|
print(result["answer"]) |
|
|
|
