Text Generation On Emnlp2017 Wmt
評価指標
BLEU-2
BLEU-3
BLEU-4
BLEU-5
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
比較表
| モデル名 | BLEU-2 | BLEU-3 | BLEU-4 | BLEU-5 |
|---|---|---|---|---|
| seqgan-sequence-generative-adversarial-nets | 0.859 | 0.6015 | 0.4541 | 0.4498 |
| improving-gan-training-with-probability-ratio | 0.905 | 0.692 | 0.47 | 0.322 |
| adversarial-ranking-for-language-generation | 0.778 | 0.478 | 0.411 | 0.463 |
| long-text-generation-via-adversarial-training | 0.956 | 0.819 | 0.627 | 0.498 |
| relgan-relational-generative-adversarial | 0.881 | 0.705 | 0.501 | 0.319 |