Closed vpratz closed 4 months ago
Very nice! The ConfigurableMLP is now part of the library in helper_networks.py
, so the notebook can directly import and use it.
Thanks for the review! I extended ConfigurableMLP
to allow for arbitrary output dimensionality, so now we can use it for the current use case (input_dim!=output_dim) as well.
Cool and good idea with the different dimensionalities!
A quick demonstration on how to use the new
AmortizedPointEstimator
class discussed in #121. Contains the illustrative example of the original paper.