asahi417's picture
model update
522248c
|
raw
history blame
7.46 kB
metadata
license: cc-by-4.0
metrics:
  - bleu4
  - meteor
  - rouge-l
  - bertscore
  - moverscore
language: fr
datasets:
  - lmqg/qag_frquad
pipeline_tag: text2text-generation
tags:
  - questions and answers generation
widget:
  - text: >-
      Créateur » (Maker), lui aussi au singulier, « le Suprême Berger » (The
      Great Shepherd) ; de l'autre, des réminiscences de la théologie de
      l'Antiquité : le tonnerre, voix de Jupiter, « Et souvent ta voix gronde en
      un tonnerre terrifiant », etc.
    example_title: Questions & Answers Generation Example 1
model-index:
  - name: lmqg/mt5-small-frquad-qag
    results:
      - task:
          name: Text2text Generation
          type: text2text-generation
        dataset:
          name: lmqg/qag_frquad
          type: default
          args: default
        metrics:
          - name: BLEU4 (Question & Answer Generation)
            type: bleu4_question_answer_generation
            value: 1.54
          - name: ROUGE-L (Question & Answer Generation)
            type: rouge_l_question_answer_generation
            value: 15.33
          - name: METEOR (Question & Answer Generation)
            type: meteor_question_answer_generation
            value: 16.12
          - name: BERTScore (Question & Answer Generation)
            type: bertscore_question_answer_generation
            value: 64.81
          - name: MoverScore (Question & Answer Generation)
            type: moverscore_question_answer_generation
            value: 50.01
          - name: QAAlignedF1Score-BERTScore (Question & Answer Generation)
            type: qa_aligned_f1_score_bertscore_question_answer_generation
            value: 77.23
          - name: QAAlignedRecall-BERTScore (Question & Answer Generation)
            type: qa_aligned_recall_bertscore_question_answer_generation
            value: 77.74
          - name: QAAlignedPrecision-BERTScore (Question & Answer Generation)
            type: qa_aligned_precision_bertscore_question_answer_generation
            value: 76.76
          - name: QAAlignedF1Score-MoverScore (Question & Answer Generation)
            type: qa_aligned_f1_score_moverscore_question_answer_generation
            value: 52.36
          - name: QAAlignedRecall-MoverScore (Question & Answer Generation)
            type: qa_aligned_recall_moverscore_question_answer_generation
            value: 52.54
          - name: QAAlignedPrecision-MoverScore (Question & Answer Generation)
            type: qa_aligned_precision_moverscore_question_answer_generation
            value: 52.19

Model Card of lmqg/mt5-small-frquad-qag

This model is fine-tuned version of google/mt5-small for question & answer pair generation task on the lmqg/qag_frquad (dataset_name: default) via lmqg.

Overview

Usage

from lmqg import TransformersQG

# initialize model
model = TransformersQG(language="fr", model="lmqg/mt5-small-frquad-qag")

# model prediction
question_answer_pairs = model.generate_qa("Créateur » (Maker), lui aussi au singulier, « le Suprême Berger » (The Great Shepherd) ; de l'autre, des réminiscences de la théologie de l'Antiquité : le tonnerre, voix de Jupiter, « Et souvent ta voix gronde en un tonnerre terrifiant », etc.")
  • With transformers
from transformers import pipeline

pipe = pipeline("text2text-generation", "lmqg/mt5-small-frquad-qag")
output = pipe("Créateur » (Maker), lui aussi au singulier, « le Suprême Berger » (The Great Shepherd) ; de l'autre, des réminiscences de la théologie de l'Antiquité : le tonnerre, voix de Jupiter, « Et souvent ta voix gronde en un tonnerre terrifiant », etc.")

Evaluation

Score Type Dataset
BERTScore 64.81 default lmqg/qag_frquad
Bleu_1 9.63 default lmqg/qag_frquad
Bleu_2 4.73 default lmqg/qag_frquad
Bleu_3 2.64 default lmqg/qag_frquad
Bleu_4 1.54 default lmqg/qag_frquad
METEOR 16.12 default lmqg/qag_frquad
MoverScore 50.01 default lmqg/qag_frquad
QAAlignedF1Score (BERTScore) 77.23 default lmqg/qag_frquad
QAAlignedF1Score (MoverScore) 52.36 default lmqg/qag_frquad
QAAlignedPrecision (BERTScore) 76.76 default lmqg/qag_frquad
QAAlignedPrecision (MoverScore) 52.19 default lmqg/qag_frquad
QAAlignedRecall (BERTScore) 77.74 default lmqg/qag_frquad
QAAlignedRecall (MoverScore) 52.54 default lmqg/qag_frquad
ROUGE_L 15.33 default lmqg/qag_frquad

Training hyperparameters

The following hyperparameters were used during fine-tuning:

  • dataset_path: lmqg/qag_frquad
  • dataset_name: default
  • input_types: ['paragraph']
  • output_types: ['questions_answers']
  • prefix_types: None
  • model: google/mt5-small
  • max_length: 512
  • max_length_output: 256
  • epoch: 13
  • batch: 8
  • lr: 0.001
  • fp16: False
  • random_seed: 1
  • gradient_accumulation_steps: 8
  • label_smoothing: 0.0

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",
}