Edit model card

Model Card of lmqg/mt5-small-koquad-qg-ae

This model is fine-tuned version of google/mt5-small for question generation and answer extraction jointly on the lmqg/qg_koquad (dataset_name: default) via lmqg.

Overview

Usage

from lmqg import TransformersQG

# initialize model
model = TransformersQG(language="ko", model="lmqg/mt5-small-koquad-qg-ae")

# model prediction
question_answer_pairs = model.generate_qa("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.")
  • With transformers
from transformers import pipeline

pipe = pipeline("text2text-generation", "lmqg/mt5-small-koquad-qg-ae")

# answer extraction
answer = pipe("generate question: 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.")

# question generation
question = pipe("extract answers: 또한 스피어스는 많은 새로운 여성 아티스트들에게 영향을 끼쳤는데, 대표적으로 데미 로바토, 케이티 페리, 크리스티니아 드바지, 레이디 가가, 리틀 부츠, 셀레나 고메즈 & 더씬, 픽시 로트 이 있다. 2007년 비욘세 놀스는 Total Request Live와의 인터뷰에서 '나는 브리트니를 사랑하고 팬이에요. 특히 새 앨범 Blackout을 좋아해요'라고 말했다. 린제이 로한은 '언제나 브리트니 스피어스에게 영감을 받는다. 학창시절 그녀처럼 타블로이드에 오르기를 꿈꿔왔다'고 말하며 롤 모델로 꼽았다. 스피어스는 현대 음악가들에게 음악적 영감으로 언급되기도 했다. <hl> 마일리 사이러스는 자신의 히트곡 Party in the U.S.A. 가 브리트니에게 영감과 영향을 받은 곡이라고 밝혔다. <hl> 베리 매닐로우의 앨범 15 Minutes 역시 브리트니에게 영감을 얻었다고 언급되었다.")

Evaluation

Score Type Dataset
BERTScore 83.4 default lmqg/qg_koquad
Bleu_1 25.91 default lmqg/qg_koquad
Bleu_2 19.09 default lmqg/qg_koquad
Bleu_3 14.37 default lmqg/qg_koquad
Bleu_4 10.91 default lmqg/qg_koquad
METEOR 27.52 default lmqg/qg_koquad
MoverScore 82.54 default lmqg/qg_koquad
ROUGE_L 25.83 default lmqg/qg_koquad
Score Type Dataset
QAAlignedF1Score (BERTScore) 80.36 default lmqg/qg_koquad
QAAlignedF1Score (MoverScore) 82.55 default lmqg/qg_koquad
QAAlignedPrecision (BERTScore) 77.34 default lmqg/qg_koquad
QAAlignedPrecision (MoverScore) 78.93 default lmqg/qg_koquad
QAAlignedRecall (BERTScore) 83.72 default lmqg/qg_koquad
QAAlignedRecall (MoverScore) 86.69 default lmqg/qg_koquad
Score Type Dataset
AnswerExactMatch 80.78 default lmqg/qg_koquad
AnswerF1Score 86.98 default lmqg/qg_koquad
BERTScore 95.65 default lmqg/qg_koquad
Bleu_1 75.14 default lmqg/qg_koquad
Bleu_2 66.16 default lmqg/qg_koquad
Bleu_3 53.61 default lmqg/qg_koquad
Bleu_4 38.2 default lmqg/qg_koquad
METEOR 59.91 default lmqg/qg_koquad
MoverScore 94.61 default lmqg/qg_koquad
ROUGE_L 82.32 default lmqg/qg_koquad

Training hyperparameters

The following hyperparameters were used during fine-tuning:

  • dataset_path: lmqg/qg_koquad
  • dataset_name: default
  • input_types: ['paragraph_answer', 'paragraph_sentence']
  • output_types: ['question', 'answer']
  • prefix_types: ['qg', 'ae']
  • model: google/mt5-small
  • max_length: 512
  • max_length_output: 32
  • epoch: 6
  • 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.

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",
}
Downloads last month
47

Dataset used to train lmqg/mt5-small-koquad-qg-ae

Evaluation results

  • BLEU4 (Question Generation) on lmqg/qg_koquad
    self-reported
    10.910
  • ROUGE-L (Question Generation) on lmqg/qg_koquad
    self-reported
    25.830
  • METEOR (Question Generation) on lmqg/qg_koquad
    self-reported
    27.520
  • BERTScore (Question Generation) on lmqg/qg_koquad
    self-reported
    83.400
  • MoverScore (Question Generation) on lmqg/qg_koquad
    self-reported
    82.540
  • QAAlignedF1Score-BERTScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    80.360
  • QAAlignedRecall-BERTScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    83.720
  • QAAlignedPrecision-BERTScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    77.340
  • QAAlignedF1Score-MoverScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    82.550
  • QAAlignedRecall-MoverScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    86.690
  • QAAlignedPrecision-MoverScore (Question & Answer Generation (with Gold Answer)) on lmqg/qg_koquad
    self-reported
    78.930
  • BLEU4 (Answer Extraction) on lmqg/qg_koquad
    self-reported
    38.200
  • ROUGE-L (Answer Extraction) on lmqg/qg_koquad
    self-reported
    82.320
  • METEOR (Answer Extraction) on lmqg/qg_koquad
    self-reported
    59.910
  • BERTScore (Answer Extraction) on lmqg/qg_koquad
    self-reported
    95.650
  • MoverScore (Answer Extraction) on lmqg/qg_koquad
    self-reported
    94.610
  • AnswerF1Score (Answer Extraction) on lmqg/qg_koquad
    self-reported
    86.980
  • AnswerExactMatch (Answer Extraction) on lmqg/qg_koquad
    self-reported
    80.780