CommonGen Leaderboard

Rank

Model

BLEU-4

METEOR

CIDEr

SPICE

1

Aug 26, 2020

KG-BART

University of Illinois at Chicago

Email   Paper
30.9 32.4 16.83 32.70

2

Oct 12, 2020

EKI-BART

MSRA and Fudan University

Email   Paper (COLING 2020)
32.5 32.4 16.84 32.40

3

Jun 1, 2020

T5-Large

Fine-tuned by USC-INK

T5 Paper
28.60 30.10 14.96 31.60

4

Jun 1, 2020

BART

Fine-tuned by USC-INK

BART Paper
26.30 30.90 13.92 30.60

5

Jun 1, 2020

UniLM

Fine-tuned by USC-INK

UniLM Paper
27.70 29.70 14.85 30.20

6

Jun 1, 2020

T5-Base

Fine-tuned by USC-INK

T5 Paper
16.40 23.00 9.16 22.00

7

Jun 1, 2020

BERT-Gen

Fine-tuned by USC-INK

Code
21.10 27.30 12.49 27.30

8

Jun 1, 2020

GPT-2

Fine-tuned by USC-INK

GPT-2 Paper
21.10 26.20 12.15 25.90

* These models were used as baseline methods in our paper, which were fine-tuned on the training data of CommonGen by us.

Submit to this leaderboard: You can submit your prediction by sending email to yuchen.lin@usc.edu with the title "CommonGen submission (your model name)" and the same format of this example prediction file.



Misc.

Citation

@article{lin20comgen,
    author = {Bill Yuchen Lin  and Wangchunshu Zhou and Ming Shen and Pei Zhou and Chandra Bhagavatula and Yejin Choi and Xiang Ren},
    title = {CommonGen: A Constrained Text Generation Challenge for Generative Commonsense Reasoning},
    journal = {Findings of EMNLP},
    year = {2020},
}