Tencent/TurboTransformers

Refactor MultiheadAttention and other Layers

Offen

#140 geöffnet am 22.07.2020

 (0 Kommentare) (0 Reaktionen) (0 zugewiesene Personen)C++ (206 Forks)github user discovery
enhancementhelp wanted

Repository-Metriken

Stars
 (1.546 Sterne)
PR-Merge-Metriken
 (PR-Metriken ausstehend)

Beschreibung

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.

Contributor Guide