Something changed in `pytensor > 2.12.3` (and thus `pymc > 5.6.1`) that makes my `pytensor.gradient.grad` call get stuck - any ideas?

Wow, this issue was an education… Thanks guys for all your help!

I think I’ll treat this thread as closed, since I’ve changed several things and finally somehow arrived at a model that includes the log_jcd and samples without complaint! I’m going to carefully extend it through the full workflow and watch for issues…