Mahé Perrette
Mahé Perrette
Thanks, that's certainly useful to know. Will work on a preemptive fix soon.
Oh yes this is a bug i'm in the process of fixing in a larger pr about git tracking (there is a branch but no pr yet). More in a...
Hmm, I wouldn't know at this point. If that fixes your issue that's good, but I wonder if overwriting HOME might have undesired side effects. I see 3.10 fails, but...
I can do that. This is a bit of an ambitious branch in terms of how the undo is handled (it sets up a git repository in a central location...
Hi @gl-yziquel , it's a good idea, thanks for the suggestion. For now it's not possible. You'd have to add a bibtex entry crafted manually or with other tools (e...
@boyanpenkov I think we are now in a better shape to work on the tests (split by thematics, coverage works). In case you wish to work on adding some, don't...
Just passing by: it does look that a sensibile behavior (sampling the weight *indices*) and a bad documentation. Definitely not a bug.
And here the more general case: ```python import numpy as np import arviz from scipy.interpolate import interp1d, InterpolatedUnivariateSpline from scipy.stats import norm, lognorm import pytensor.tensor as pt import pymc as...
Thanks for paying attention and for laying out where what could go. Unfortunately at the moment I don't have the resources (time) to invest in proper implementation (I deleted by...
I could check in my real-world use-case that the approach is working, which is the satisfying part. Unfortunately further checks on better machines still yield a longer sampling time for...