numpyro
numpyro copied to clipboard
pyro-ppl
Probabilistic programming with NumPy powered by JAX for autograd and JIT compilation to GPU/TPU/CPU.
### Bug Description This is a part of the issues reported in https://github.com/pyro-ppl/numpyro/issues/1981. Running the following test will raise an error/xfail. ### Steps to Reproduce ``` JAX_CHECK_TRACER_LEAKS=1 pytest -vs test/contrib/test_infer_discrete.py::test_scan_hmm_smoke...
### Feature Summary Recently, a couple of samplers have been proposed by the group I work with (https://arxiv.org/abs/2503.01707, https://arxiv.org/abs/2212.08549) as alternatives to NUTS HMC (see issue #1662 ). Since they...
### Bug Description In the PR, adding type hints to the `distributions` module https://github.com/pyro-ppl/numpyro/pull/2032, the CI did not check for consistency, and we have many lint errors. ### Steps to...
Fixes #1998 The current `infer_discrete` tests pass so I believe the change is equivalent.
This PR unifies constraint handling for several auto guides. In short, distributions for all variables in the guide are constructed in unconstrained space. If the variable has non-real support, the...
Implementation proposal for [Issue #2030]. So far all unit tests are passing (including documentation tests). I have not implemented the CDF (also not available in the most recent version of...
### Feature Summary Adding an implementation of the Beta Negative Binomial distribution to the conjugate.py module. https://en.wikipedia.org/wiki/Beta_negative_binomial_distribution ### Why is this needed? Provides a parametric heavy tailed count distribution. The...
### Feature Summary The Inverse Wishart Distribution should be added into the distribution class. Then it is more convenient to directly use it instead of computing the inverse of a...
### Bug Description `TraceEnum_ELBO` ignores the `is_auxiliary` option of `infer` in sampling statements. ### Steps to Reproduce Apologies, these code snippets are a bit long. #### Model ``` def mix_weights(beta):...
### Bug Description plate_stack doesn't work as expected. I'll describe the issue with an example below: ### Steps to Reproduce In the code below, in the _model callable, I want...
