Edit model card

Model Card of lmqg/mbart-large-cc25-frquad-qg

This model is fine-tuned version of facebook/mbart-large-cc25 for question generation task on the lmqg/qg_frquad (dataset_name: default) via lmqg.

Overview

Usage

from lmqg import TransformersQG

# initialize model
model = TransformersQG(language="fr", model="lmqg/mbart-large-cc25-frquad-qg")

# model prediction
questions = model.generate_q(list_context="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.", list_answer="le Suprême Berger")
  • With transformers
from transformers import pipeline

pipe = pipeline("text2text-generation", "lmqg/mbart-large-cc25-frquad-qg")
output = pipe("Créateur » (Maker), lui aussi au singulier, « <hl> le Suprême Berger <hl> » (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 81.75 default lmqg/qg_frquad
Bleu_1 30.64 default lmqg/qg_frquad
Bleu_2 19.09 default lmqg/qg_frquad
Bleu_3 13.26 default lmqg/qg_frquad
Bleu_4 9.47 default lmqg/qg_frquad
METEOR 19.8 default lmqg/qg_frquad
MoverScore 57.96 default lmqg/qg_frquad
ROUGE_L 30.62 default lmqg/qg_frquad
  • Metric (Question & Answer Generation, Reference Answer): Each question is generated from the gold answer. raw metric file
Score Type Dataset
QAAlignedF1Score (BERTScore) 81.27 default lmqg/qg_frquad
QAAlignedF1Score (MoverScore) 55.61 default lmqg/qg_frquad
QAAlignedPrecision (BERTScore) 81.29 default lmqg/qg_frquad
QAAlignedPrecision (MoverScore) 55.61 default lmqg/qg_frquad
QAAlignedRecall (BERTScore) 81.25 default lmqg/qg_frquad
QAAlignedRecall (MoverScore) 55.6 default lmqg/qg_frquad
Score Type Dataset
QAAlignedF1Score (BERTScore) 75.55 default lmqg/qg_frquad
QAAlignedF1Score (MoverScore) 51.75 default lmqg/qg_frquad
QAAlignedPrecision (BERTScore) 74.04 default lmqg/qg_frquad
QAAlignedPrecision (MoverScore) 51.03 default lmqg/qg_frquad
QAAlignedRecall (BERTScore) 77.16 default lmqg/qg_frquad
QAAlignedRecall (MoverScore) 52.52 default lmqg/qg_frquad

Training hyperparameters

The following hyperparameters were used during fine-tuning:

  • dataset_path: lmqg/qg_frquad
  • dataset_name: default
  • input_types: ['paragraph_answer']
  • output_types: ['question']
  • prefix_types: None
  • model: facebook/mbart-large-cc25
  • max_length: 512
  • max_length_output: 32
  • epoch: 8
  • batch: 4
  • lr: 0.001
  • fp16: False
  • random_seed: 1
  • gradient_accumulation_steps: 16
  • 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
4

Evaluation results

  • BLEU4 (Question Generation) on lmqg/qg_frquad
    self-reported
    9.470
  • ROUGE-L (Question Generation) on lmqg/qg_frquad
    self-reported
    30.620
  • METEOR (Question Generation) on lmqg/qg_frquad
    self-reported
    19.800
  • BERTScore (Question Generation) on lmqg/qg_frquad
    self-reported
    81.750
  • MoverScore (Question Generation) on lmqg/qg_frquad
    self-reported
    57.960
  • QAAlignedF1Score-BERTScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    81.270
  • QAAlignedRecall-BERTScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    81.250
  • QAAlignedPrecision-BERTScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    81.290
  • QAAlignedF1Score-MoverScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    55.610
  • QAAlignedRecall-MoverScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    55.600
  • QAAlignedPrecision-MoverScore (Question & Answer Generation (with Gold Answer)) [Gold Answer] on lmqg/qg_frquad
    self-reported
    55.610
  • QAAlignedF1Score-BERTScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    75.550
  • QAAlignedRecall-BERTScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    77.160
  • QAAlignedPrecision-BERTScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    74.040
  • QAAlignedF1Score-MoverScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    51.750
  • QAAlignedRecall-MoverScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    52.520
  • QAAlignedPrecision-MoverScore (Question & Answer Generation) [Gold Answer] on lmqg/qg_frquad
    self-reported
    51.030