ONE - On-device Neural Engine
|
#include <IGradientApplier.h>
Public Member Functions | |
virtual | ~IGradientApplier ()=default |
virtual void | applyGradient (uint32_t training_step)=0 |
Apply gradients to a trainable tensor. | |
Definition at line 29 of file IGradientApplier.h.
|
virtualdefault |
|
pure virtual |
Apply gradients to a trainable tensor.
training_step | The number of iterations of the training process. |
Implemented in onert::backend::train::ops::GradientApplier.