I have a log-prob function and its gradient. Because FFT is involved in the log-prob function and its gradient, it is not possible to perform autodiff. So I do the differential by hand. Is there a (documented) way to call the NUTS sampler directly? I did notice there is a way to create a customize…

The recommanded approach is create a custom Theano Ops with its gradient. You can see a example here: https://docs.pymc.io/Advanced_usage_of_Theano_in_PyMC3.html#writing-custom-theano-ops and a more in complex example here: https://docs.pymc.io/notebooks/blackbox_external_likelihood.html

Is there a way to call NUTS directly to sample a customized likelihood together with its gradient?

Questions

nkaimcaudle March 13, 2020, 9:51am 3

I am not sure but perhaps this is what you are after Isolate NUTS into a new library

Topic		Replies	Views
Auto-diff when using NUTS sampler with blackbox likelihoods Questions development	1	868	September 11, 2020
Black box NUTS sampler with multidimensional parameters v5 modeling	7	164	April 7, 2025
Using pymc3 NUTS sampler for an "external model" having its derivatives with respect to its parameters Questions	4	2421	March 30, 2018
NUTS - FromFunctionOp has no attribute 'grad' Questions theano	2	1026	August 3, 2018
NUTS can sample even failing to differentiate? Questions	6	838	September 23, 2018

Is there a way to call NUTS directly to sample a customized likelihood together with its gradient?

Related topics