Say we already have define
a = mx.sym.Variable('a')
b = mx.sym.Variable('b')
c = a+b # name is _plus0_output
d = c/2
and we could calculate c
with
c.eval(a=mx.nd.array([2.0]),b=mx.nd.array([3.0]))
or
d.get_internals().eval(a=mx.nd.array([2.0]),b=mx.nd.array([3.0]))
to get both c
and d
.
but how could I calculate d
only via c
?
sometimes I’d like to use synthetic c
to check if the model is robust and in this case I could not get a
and b
.
I’d like to have something like
d.eval(_plus0_output=mx.nd.array([2.5]))
Is it possible?