| Paper | Code | BLEU-2 | BLEU-3 | BLEU-4 | BLEU-5 | NLLgen | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|---|---|
| Long Text Generation via Adversarial Training with Leaked Information | ✓ Link | 0.956 | 0.819 | 0.627 | 0.498 | LeakGAN | 2017-09-24 | |
| Improving GAN Training with Probability Ratio Clipping and Sample Reweighting | ✓ Link | 0.905 | 0.692 | 0.47 | 0.322 | 2.265 | PPOGAN | 2020-06-12 |
| RelGAN: Relational Generative Adversarial Networks for Text Generation | ✓ Link | 0.881 | 0.705 | 0.501 | 0.319 | RelGAN | 2019-05-01 | |
| SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient | ✓ Link | 0.859 | 0.6015 | 0.4541 | 0.4498 | SeqGAN | 2016-09-18 | |
| Adversarial Ranking for Language Generation | ✓ Link | 0.778 | 0.478 | 0.411 | 0.463 | RankGAN | 2017-05-31 |