I’m looking for a good example of a model failing mid-run. Some context: I’m working on updating a “How to debug a model” example from pymc-examples. The old example doesn’t work anymore because pymc exception messages got so much better (it makes aesara.printing.Print unnecessary - which is awesome by the way). I played with various badly defined models (things like sigma = a+b*x, Neal’s funnel), but I can’t get my models to fail mid-run. Sure, they produce terrible diagnostics (like high R-hat and low ess), but in those cases I can just route problematic intermediate values to a Deterministic variable (which again removes the need for aesara.printing.Print). I can use one of these models as an example, but it doesn’t seem like a good solution. Any thoughts? Thanks!
2 Likes
Hey there @Igor_K, did you ever find a good example of aesara.printing.Print? I’m currently trying to use it also and I don’t seem to be having much success. Thanks!