def function(self, *args: tp.Any, **kwargs: tp.Any) -> tp.Loss:
out = self._func(*args, **kwargs)
if isinstance(out, (bool, np.bool_)):
raise errors.NevergradTypeError(
"Constraint must be a positive float if unsatisfied constraint (not bool)"
)
return np.maximum(0, out) # type: ignore
def __init__(self, *args: tp.Any, **kwargs: tp.Any) -> None:
super().__init__()
if any(
isinstance(x, Parameter)
for x in args + tuple(kwargs.values())):
raise errors.NevergradTypeError(
"Operation with Parameter instances are not supported")
def _layered_sample(self) -> Data:
if not self.uniform_sampling:
return super()._layered_sample() # type: ignore
root = self._layers[0]
if not isinstance(root, Data):
raise errors.NevergradTypeError(f"BoundLayer {self} on a non-Data root {root}")
shape = super()._layered_get_value().shape
child = root.spawn_child()
# send new val to the layer under this one for the child
new_val = self.random_state.uniform(size=shape)
del child.value # make sure there is no cache
child._layers[self._layer_index].set_normalized_value(new_val) # type: ignore
return child