File size: 1,626 Bytes
f1bb776 898873c d2f814d f1bb776 898873c d2f814d 898873c d2f814d 898873c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 |
---
license: mit
datasets:
- McGill-NLP/FaithDial
widget:
- text: "A cardigan is a type of knitted garment (sweater) that has an open front. </s></s> The old version is the regular one, knitted garment that has open front and buttons!"
---
## Overview
FaithCritic is the [RoBERTa large model](https://huggingface.co/roberta-large) fine-tuned on the [FaithDial](https://huggingface.co/datasets/McGill-NLP/FaithDial) dataset. The objective is to predict whether an utterance is faithful or not, given the source knowledge.
The hyperparameters are provided in [hparams.yml](https://huggingface.co/McGill-NLP/roberta-large-faithcritic/blob/main/hparams.yaml). To know more about how to train a critic model, visit [our repo](https://github.com/McGill-NLP/FaithDial).
## Usage
```python
from transformers import AutoTokenizer, AutoModel
tokenizer = AutoTokenizer.from_pretrained("McGill-NLP/roberta-large-faithcritic")
model = AutoModel.from_pretrained("McGill-NLP/roberta-large-faithcritic")
knowledge = "A cardigan is a type of knitted garment (sweater) that has an open front."
response = "The old version is the regular one, knitted garment that has open front and buttons!"
input = tokenizer(knowledge, response)
output = model(**input)
```
## Citation Information
```bibtex
@article{dziri2022faithdial,
title={FaithDial: A Faithful Benchmark for Information-Seeking Dialogue},
author={Dziri, Nouha and Kamalloo, Ehsan and Milton, Sivan and Zaiane, Osmar and Yu, Mo and Ponti, Edoardo and Reddy, Siva},
journal={arXiv preprint, arXiv:2204.10757},
year={2022},
url={https://arxiv.org/abs/2204.10757}
}
```
|