Hard-Coded Gaussian Attention for Neural Machine Translation
37285Knowledge Triples
| Source Entity | Relationship | Target Entity |
|---|---|---|
|
Трансформер
Method |
comparesWith |