Repository navigation
Conditional sampling with vmap #157
Description
Activity
Another unclear thing for me with
plateandvmapusage in #39 is when I have a different number of observations for each group in a hierarchical model (e.g. lightcurves with different number of data points for different objects). In that case usingnumpyro.handlers.maskis suggested here. But this implies using the secondplate(per-observation). Is there any workaround?
Hi!
As discussed in #39 one has to use
vmapwith numpyroplate. But how to obtain a conditional samples (e.g. on a finer X grid) in that case? In this tutorial it is done as follows (but w/o vmaping):numpyro.deterministic("pred", gp.condition(y, true_t).gp.loc)However, I can't get
GaussianProcessinstance out of thevmap-ed functionbuild_gp(as in #39).Thanks in advance!
Ilya