Improving Transformer Models by Reordering their Sublayers
37902Knowledge Triples
| Source Entity | Relationship | Target Entity |
|---|---|---|
|
Multilayer Transformer Networks
OtherEntity |
includes |
Feedforward Sublayers
Method |
|
Multilayer Transformer Networks
OtherEntity |
includes |
Self Attention
Method |
|
Sandwich Transformer
Method |
improves |
Perplexity
Metric |