pymc
pymc copied to clipboard
Bayesian Modeling and Probabilistic Programming in Python
### Description Rely on PyTensor 3.0, with default numba backend. This makes PyMC safely pip installable Potentially could make nutpie the default sampler when the whole model can run with...
## Description ## Changes - Added `icdf` implementation to `pymc/distributions/discrete.py`. - Added `test_categorical_icdf`, `test_categorical_icdf_batch_shapes`, and `test_categorical_icdf_upper_edge`. ## Testing - All Categorical tests pass locally (`pytest -q tests/distributions -k categorical`). -...
### Issue with current documentation: I tried out the ZarrTrace backend and didn't notice anything being written. A directory is created when I make the `ZarrTrace` object, but it isn't...
## Description - Add abstract transform_coords method to Transform base class - Implement method in all concrete Transform classes - SimplexTransform and SumTo1 remove last coordinate - Most other transforms...
## Description We will implement `ICARRV` and `CARRV` as `SymbolicRV`s by subclassing `SymbolicMVNormalUsedInternally`. This will eliminate the need for `rng_fn`. ## Related Issue - [x] Closes #7713 - [ ]...
- [x] All arviz plots must be retrieved from pymc.plots.foo - [x] All arviz stats must be retrieved from pymc.stats.foo - [x] Removed many rare or non-user facing objects -...
## Description Initial work on #7462 from the PyData Amsterdam 2025 PyMC sprint. ---- ๐ Documentation preview ๐: https://pymc--7918.org.readthedocs.build/en/7918/
### Describe the issue: Following these docs on the contributing page: ``` conda env create -f conda-envs/environment-dev.yml ``` and attempting to run the test suite results in `ModuleNotFoundError`s due to...
### Description This is a painful gotcha I get hit by once in a while. If you have inconsistencies between the model coords and variable shapes, PyMC will still happy...
In https://github.com/pymc-devs/pymc/pull/7680#issuecomment-2709633765 it was suggested that instead of deprecating `find_constrained_prior`, we should call PreliZ's maxent function. ---- ๐ Documentation preview ๐: https://pymc--7905.org.readthedocs.build/en/7905/