模型:
lmqg/t5-base-squad-qg
这个模型是通过在数据集 default 上使用 lmqg 进行问题生成任务而微调的 t5-base 的版本。
from lmqg import TransformersQG # initialize model model = TransformersQG(language="en", model="lmqg/t5-base-squad-qg") # model prediction questions = model.generate_q(list_context="William Turner was an English painter who specialised in watercolour landscapes", list_answer="William Turner")
from transformers import pipeline pipe = pipeline("text2text-generation", "lmqg/t5-base-squad-qg") output = pipe("generate question: <hl> Beyonce <hl> further expanded her acting career, starring as blues singer Etta James in the 2008 musical biopic, Cadillac Records.")
Score | Type | Dataset | |
---|---|---|---|
BERTScore | 90.6 | default | 12313321 |
Bleu_1 | 58.69 | default | 12313321 |
Bleu_2 | 42.66 | default | 12313321 |
Bleu_3 | 32.99 | default | 12313321 |
Bleu_4 | 26.13 | default | 12313321 |
METEOR | 26.97 | default | 12313321 |
MoverScore | 64.74 | default | 12313321 |
ROUGE_L | 53.33 | default | 12313321 |
Score | Type | Dataset | |
---|---|---|---|
QAAlignedF1Score (BERTScore) | 95.42 | default | 12313321 |
QAAlignedF1Score (MoverScore) | 70.63 | default | 12313321 |
QAAlignedPrecision (BERTScore) | 95.48 | default | 12313321 |
QAAlignedPrecision (MoverScore) | 70.92 | default | 12313321 |
QAAlignedRecall (BERTScore) | 95.37 | default | 12313321 |
QAAlignedRecall (MoverScore) | 70.34 | default | 12313321 |
Score | Type | Dataset | |
---|---|---|---|
QAAlignedF1Score (BERTScore) | 92.75 | default | 12313321 |
QAAlignedF1Score (MoverScore) | 64.36 | default | 12313321 |
QAAlignedPrecision (BERTScore) | 92.59 | default | 12313321 |
QAAlignedPrecision (MoverScore) | 64.45 | default | 12313321 |
QAAlignedRecall (BERTScore) | 92.93 | default | 12313321 |
QAAlignedRecall (MoverScore) | 64.35 | default | 12313321 |
Dataset | Type | BERTScore | Bleu_4 | METEOR | MoverScore | ROUGE_L | Link |
---|---|---|---|---|---|---|---|
12336321 | amazon | 90.75 | 6.57 | 22.37 | 60.8 | 24.81 | 12337321 |
12336321 | new_wiki | 93.02 | 11.09 | 27.23 | 65.97 | 29.59 | 12339321 |
12336321 | nyt | 92.2 | 7.77 | 25.16 | 63.83 | 24.56 | 12341321 |
12336321 | 90.59 | 5.68 | 21.3 | 60.23 | 21.96 | 12343321 | |
12344321 | books | 88.14 | 0.49 | 13.51 | 55.65 | 9.44 | 12345321 |
12344321 | electronics | 87.71 | 0.0 | 16.53 | 55.77 | 13.48 | 12347321 |
12344321 | grocery | 87.46 | 0.0 | 16.24 | 56.59 | 10.26 | 12349321 |
12344321 | movies | 87.66 | 0.72 | 13.06 | 55.45 | 11.89 | 12351321 |
12344321 | restaurants | 87.83 | 0.0 | 13.3 | 55.45 | 10.7 | 12353321 |
12344321 | tripadvisor | 89.23 | 0.93 | 16.51 | 56.67 | 13.51 | 12355321 |
在微调过程中使用了以下超参数:
可以在 fine-tuning config file 找到完整的配置。
@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", }