Example: CO2 at Mauna Loa gives wrong results


I am trying to run this example just by copy-pasting the code pieces into VSC: Example: CO2 at Mauna Loa — PyMC3 3.11.4 documentation

However I get absurd hyperparameters as result:
[216/216 02:55<00:00 logp = 1,730.7, ||grad|| = 3.9171]

[‘period:1.2375350961999776’, ‘α:3255.948700168235’, ‘η_med:1090880291.369142’, ‘η_noise:8.592077946469366e-14’, ‘η_per:6.2426404996515864e-24’, ‘η_trend:7.5692537078421705e-06’, ‘σ:0.00029984624804483184’, ‘ℓ_med:1.9937404288132644e-07’, ‘ℓ_noise:386.1405617864199’, ‘ℓ_pdecay:4.4276092788725187e-11’, ‘ℓ_psmooth :0.03312052370660301’, ‘ℓ_trend:2.2809154610959933e+26’]

So is it possible to download the corresponding working jupiter notebook somewhere?

All the examples can be found in the pymc examples repo. There appear to be 2 notebooks for the Mauna Loa GP example? Not sure why. But you can check them out here and here.

Edit: You can also find the up-to-date render of that notebook (or one of those notebooks) in the new pymc3 examples gallery found here.


Thank you Christian, this is what I was looking for, I will try to run it om my own machine.