---
language:
- en
license: cc-by-4.0
datasets:
- cuad
pipeline_tag: question-answering
tags:
- legal-contract-review
- roberta
- cuad
library_name: transformers
---
# Model Card for cuad-roberta-base
  
# Model Details
 
## Model Description
 
- **Developed by:** Hendrycks et al.
- **Model type:** Question Answering 
- **Language(s) (NLP):** en
- **License:** cc-by-4.0
- **Related Models:**
  - **Parent Model:** DeBERTa-v2 
- **Resources for more information:** 
    - GitHub Repo: [TheAtticusProject](https://github.com/TheAtticusProject/cuad)
    - Associated Paper: [CUAD: An Expert-Annotated NLP Dataset for Legal Contract Review](https://arxiv.org/abs/2103.06268)
    - Project website: [Contract Understanding Atticus Dataset (CUAD)](https://www.atticusprojectai.org/cuad)
 
# Uses
 
## Direct Use
 
This model can be used for the task of Question Answering on Legal Documents.
 
# Training Details

Read: [CUAD: An Expert-Annotated NLP Dataset for Legal Contract Review](https://arxiv.org/abs/2103.06268) 
for detailed information on training procedure, dataset preprocessing and evaluation.
 
## Training Data, Procedure, Preprocessing, etc.
 
See [CUAD dataset card](https://huggingface.co/datasets/cuad) for more information.
 
# Evaluation
 
## Testing Data, Factors & Metrics
 
### Testing Data
 
See [CUAD dataset card](https://huggingface.co/datasets/cuad) for more information.
 
### Software

Python, Transformers
 
# Citation
 
**BibTeX:**
 ```
@article{hendrycks2021cuad,
      title={CUAD: An Expert-Annotated NLP Dataset for Legal Contract Review}, 
      author={Dan Hendrycks and Collin Burns and Anya Chen and Spencer Ball},
      journal={NeurIPS},
      year={2021}
}
```
 
# How to Get Started with the Model
 
Use the code below to get started with the model.
 
<details>
<summary> Click to expand </summary>

```python
from transformers import AutoTokenizer, AutoModelForQuestionAnswering
 
tokenizer = AutoTokenizer.from_pretrained("mgigena/cuad-deberta-v2-xlarge")
 
model = AutoModelForQuestionAnswering.from_pretrained("mgigena/cuad-deberta-v2-xlarge")
```
</details>