Tencent/TurboTransformers
Refactor MultiheadAttention and other Layers
Aperta
#140 aperta il 22 lug 2020
enhancementhelp wanted
Metriche repository
- Star
- (1546 stelle)
- Metriche merge PR
- (Metriche PR in attesa)
Descrizione
Now the logic inside MultiheadAttention Layer is too complex for development. Moreover, some bugs exist in intermediate management. It is the first priority to rewrite these codes to make others easily understand what Turbo is doing.