pyjuice.parameterizations
A parameterization is the rule a sum node’s effective parameters follow: the shared, EM-trained parameters modified
by per-sample tensors – the external parameters – supplied at call time through
pc(x, sum_external_params = {ns: tensors}). A sum node is given one with
pyjuice.summate(..., parameterization = BlockScale(...)).
Base class of the external parameterizations of a sum node. |
|
A per-sample multiplicative gate on each parameter block, supplied externally at call time. |
|
A per-sample multiplicative low-rank scale on the sum node's parameters, supplied externally at call time. |