asahi417's picture
model update
f1f65ed
metadata
license: cc-by-4.0
metrics:
  - bleu4
  - meteor
  - rouge-l
  - bertscore
  - moverscore
language: en
datasets:
  - lmqg/qg_squad
pipeline_tag: text2text-generation
tags:
  - question generation
widget:
  - text: >-
      generate question: <hl>  Beyonce further expanded her acting career,
      starring as blues singer Etta James in the 2008 musical biopic, Cadillac
      Records. <hl>
    example_title: Question Generation Example 1
  - text: >-
      generate question: <hl> Beyonce further expanded her acting career,
      starring as blues singer Etta James in the 2008 musical biopic, Cadillac
      Records. <hl>
    example_title: Question Generation Example 2
  - text: >-
      generate question: <hl> Beyonce further expanded her acting career,
      starring as blues singer Etta James in the 2008 musical biopic, Cadillac
      Records . <hl>
    example_title: Question Generation Example 3
model-index:
  - name: research-backup/t5-small-squad-qg-no-answer
    results:
      - task:
          name: Text2text Generation
          type: text2text-generation
        dataset:
          name: lmqg/qg_squad
          type: default
          args: default
        metrics:
          - name: BLEU4 (Question Generation)
            type: bleu4_question_generation
            value: 21.12
          - name: ROUGE-L (Question Generation)
            type: rouge_l_question_generation
            value: 47.47
          - name: METEOR (Question Generation)
            type: meteor_question_generation
            value: 23.38
          - name: BERTScore (Question Generation)
            type: bertscore_question_generation
            value: 89.64
          - name: MoverScore (Question Generation)
            type: moverscore_question_generation
            value: 62.07

Model Card of research-backup/t5-small-squad-qg-no-answer

This model is fine-tuned version of t5-small for question generation task on the lmqg/qg_squad (dataset_name: default) via lmqg. This model is fine-tuned without answer information, i.e. generate a question only given a paragraph (note that normal model is fine-tuned to generate a question given a pargraph and an associated answer in the paragraph).

Overview

Usage

from lmqg import TransformersQG

# initialize model
model = TransformersQG(language="en", model="research-backup/t5-small-squad-qg-no-answer")

# model prediction
questions = model.generate_q(list_context="William Turner was an English painter who specialised in watercolour landscapes", list_answer="William Turner")
  • With transformers
from transformers import pipeline

pipe = pipeline("text2text-generation", "research-backup/t5-small-squad-qg-no-answer")
output = pipe("generate question: <hl>  Beyonce further expanded her acting career, starring as blues singer Etta James in the 2008 musical biopic, Cadillac Records. <hl>")

Evaluation

Score Type Dataset
BERTScore 89.64 default lmqg/qg_squad
Bleu_1 53.37 default lmqg/qg_squad
Bleu_2 36.67 default lmqg/qg_squad
Bleu_3 27.4 default lmqg/qg_squad
Bleu_4 21.12 default lmqg/qg_squad
METEOR 23.38 default lmqg/qg_squad
MoverScore 62.07 default lmqg/qg_squad
ROUGE_L 47.47 default lmqg/qg_squad

Training hyperparameters

The following hyperparameters were used during fine-tuning:

  • dataset_path: lmqg/qg_squad
  • dataset_name: default
  • input_types: ['paragraph_sentence']
  • output_types: ['question']
  • prefix_types: ['qg']
  • model: t5-small
  • max_length: 512
  • max_length_output: 32
  • epoch: 7
  • batch: 64
  • lr: 0.0001
  • 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",
}