Text Generation on EMNLP2017 WMT

Metric: BLEU-3 (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

Sort:

#	Model↕	BLEU-3▼	Extra Data	Paper	Date↕	Code
1	LeakGAN	0.819	No	Long Text Generation via Adversarial Training wi...	2017-09-24	Code
2	RelGAN	0.705	No	-	-	Code
3	PPOGAN	0.692	No	Improving GAN Training with Probability Ratio Cl...	2020-06-12	Code
4	SeqGAN	0.6015	No	SeqGAN: Sequence Generative Adversarial Nets wit...	2016-09-18	Code
5	RankGAN	0.478	No	Adversarial Ranking for Language Generation	2017-05-31	Code

#1LeakGANSOTA
0.819
BLEU-3· 2017-09-24
Long Text Generation via Adversarial Training with Leaked Information Code
#2RelGAN
0.705
BLEU-3
No paperCode
#3PPOGAN
0.692
BLEU-3· 2020-06-12
Improving GAN Training with Probability Ratio Clipping and Sample Reweighting Code
#4SeqGANSOTA
0.6015
BLEU-3· 2016-09-18
SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient Code
#5RankGAN
0.478
BLEU-3· 2017-05-31
Adversarial Ranking for Language Generation Code