jnsbck
jnsbck
I did a bit more work on this PR, current tests should be passing and I have swapped out all the legacy `CategoricalNet` code for the `CategoricalMADE`. See changes and...
Thanks for all the input
Turns out, since the posterior is an `MCMCPosterior`, only `log_prob` is used for sampling and not `sample`. This bug still eludes me. I have spent some time today and been...
Thanks a ton to both of you! I also did a bit more digging, but apart from what @janfb also found, I have nothing conclusive yet.
Thank you all for taking so much interest in this. This bug was almost literally causing me so many headaches over the last few weeks! And thanks for already making...
Amazing!
I did it :) Ready for final send off from my side. Tests passing on my end. Added a comment about option to run with >1D discrete conditions and changed...
Done :) Thanks for the final check up!
It'd be awesome if we could get this merged rather soon, since #989 depends on this, which I would love to work on. :)
awesome! thanks for the effort! :rocket: