mt5-small-koquad-qg / README.md
asahi417's picture
model update
6db5ec5
|
raw
history blame
5.61 kB
---
license: cc-by-4.0
metrics:
- bleu4
- meteor
- rouge-l
- bertscore
- moverscore
language: ko
datasets:
- lmqg/qg_koquad
pipeline_tag: text2text-generation
tags:
- question generation
widget:
- text: "1990년 영화 《 <hl> 남부군 <hl> 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다."
example_title: "Question Generation Example 1"
- text: "백신이 없기때문에 예방책은 <hl> 살충제 <hl> 를 사용하면서 서식 장소(찻찬 받침, 배수로, 고인 물의 열린 저장소, 버려진 타이어 등)의 수를 줄임으로써 매개체를 통제할 수 있다."
example_title: "Question Generation Example 2"
- text: "<hl> 원테이크 촬영 <hl> 이기 때문에 한 사람이 실수를 하면 처음부터 다시 찍어야 하는 상황이 발생한다."
example_title: "Question Generation Example 3"
model-index:
- name: lmqg/mt5-small-koquad
results:
- task:
name: Text2text Generation
type: text2text-generation
dataset:
name: lmqg/qg_koquad
type: default
args: default
metrics:
- name: BLEU4
type: bleu4
value: 0.10570915349557093
- name: ROUGE-L
type: rouge-l
value: 0.2564353531078813
- name: METEOR
type: meteor
value: 0.2752329744142515
- name: BERTScore
type: bertscore
value: 0.8288608218241639
- name: MoverScore
type: moverscore
value: 0.8249013345139385
---
# Model Card of `lmqg/mt5-small-koquad`
This model is fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) for question generation task on the
[lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) (dataset_name: default) via [`lmqg`](https://github.com/asahi417/lm-question-generation).
Please cite our paper if you use the model ([https://arxiv.org/abs/2210.03992](https://arxiv.org/abs/2210.03992)).
```
@inproceedings{ushio-etal-2022-generative,
title = "{G}enerative {L}anguage {M}odels for {P}aragraph-{L}evel {Q}uestion {G}eneration",
author = "Ushio, Asahi and
Alva-Manchego, Fernando and
Camacho-Collados, Jose",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, U.A.E.",
publisher = "Association for Computational Linguistics",
}
```
### Overview
- **Language model:** [google/mt5-small](https://huggingface.co/google/mt5-small)
- **Language:** ko
- **Training data:** [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) (default)
- **Online Demo:** [https://autoqg.net/](https://autoqg.net/)
- **Repository:** [https://github.com/asahi417/lm-question-generation](https://github.com/asahi417/lm-question-generation)
- **Paper:** [https://arxiv.org/abs/2210.03992](https://arxiv.org/abs/2210.03992)
### Usage
- With [`lmqg`](https://github.com/asahi417/lm-question-generation#lmqg-language-model-for-question-generation-)
```python
from lmqg import TransformersQG
# initialize model
model = TransformersQG(language='ko', model='lmqg/mt5-small-koquad')
# model prediction
question = model.generate_q(list_context=["1990년 영화 《 남부군 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다."], list_answer=["남부군"])
```
- With `transformers`
```python
from transformers import pipeline
# initialize model
pipe = pipeline("text2text-generation", 'lmqg/mt5-small-koquad')
# question generation
question = pipe('1990년 영화 《 <hl> 남부군 <hl> 》에서 단역으로 영화배우 첫 데뷔에 이어 같은 해 KBS 드라마 《지구인》에서 단역으로 출연하였고 이듬해 MBC 《여명의 눈동자》를 통해 단역으로 출연하였다.')
```
## Evaluation Metrics
### Metrics
| Dataset | Type | BLEU4 | ROUGE-L | METEOR | BERTScore | MoverScore | Link |
|:--------|:-----|------:|--------:|-------:|----------:|-----------:|-----:|
| [lmqg/qg_koquad](https://huggingface.co/datasets/lmqg/qg_koquad) | default | 0.106 | 0.256 | 0.275 | 0.829 | 0.825 | [link](https://huggingface.co/lmqg/mt5-small-koquad/raw/main/eval/metric.first.sentence.paragraph_answer.question.lmqg_qg_koquad.default.json) |
## Training hyperparameters
The following hyperparameters were used during fine-tuning:
- dataset_path: lmqg/qg_koquad
- dataset_name: default
- input_types: ['paragraph_answer']
- output_types: ['question']
- prefix_types: None
- model: google/mt5-small
- max_length: 512
- max_length_output: 32
- epoch: 7
- batch: 16
- lr: 0.001
- fp16: False
- random_seed: 1
- gradient_accumulation_steps: 4
- label_smoothing: 0.15
The full configuration can be found at [fine-tuning config file](https://huggingface.co/lmqg/mt5-small-koquad/raw/main/trainer_config.json).
## Citation
```
@inproceedings{ushio-etal-2022-generative,
title = "{G}enerative {L}anguage {M}odels for {P}aragraph-{L}evel {Q}uestion {G}eneration",
author = "Ushio, Asahi and
Alva-Manchego, Fernando and
Camacho-Collados, Jose",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, U.A.E.",
publisher = "Association for Computational Linguistics",
}
```