I was reading around and couldn’t figure it out.
Is there is a way to Map multiple inputs in parallel, to a module that expects a single input?
Specifically with the following scenario in mind:
def forward(self, single_x)
def forward(self, x)
x = OtherLayer(x)
output = map(MyCustomModule, x)
Is it possible to define such a thing where the feeding of the custom module will happen in parallel?
Or alternatively with a torch custom function?