You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This code implements an LSTM forward pass in CUDA using cuBLAS. It uses cuRand to initialize the inputs and parameters.
4
+
5
+
The PERFOPTS define allows the user to select which optimizations are enabled. It is a bitmask taking values from 0 to 31 with each bit defining whether a particular optimization is enabled.
0 commit comments