Sentiment Transfer on Yelp (test)
97.53Sentiment AccuracyM&M LM
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| M&M LMenergy constraint=Discriminator ↑2022.03 | 97.53 | — | 163.5 | 15.75 | 0.93 | 2.84 | |
| M&M LMenergy constraint=Hamming↑2022.03 | 94.72 | — | 191.5 | 19.71 | 0.95 | 1.83 | |
| Neural Text Attribute Transfer2017.11 | 94.4 | 77.1 | 80.1 | — | — | — | |
| Shen et al.2017.11 | 86.5 | 38.3 | 27 | — | — | — | |
| He et al.2022.03 | 84.87 | — | 200.6 | 18.67 | 0.93 | 4.23 | |
| UNMT2022.03 | 84.87 | — | 171.8 | 17 | 0.94 | 3.67 | |
| Reference Text2022.03 | 83.7 | — | 169.5 | 100 | 1 | 5.8 |