numpyro
numpyro copied to clipboard
pyro-ppl
Probabilistic programming with NumPy powered by JAX for autograd and JIT compilation to GPU/TPU/CPU.
Based on the discussion with @fritzo and @eb8680, it seems that the simplest way to support multithreading is to get PYRO_STACK from an intermediate function `get_pyro_stack`, where we can modify...
I've prepared a simple working example of Bayesian Quantile Regression using an implementation of an AsymmetricLaplace distribution in a Colab Notebook [here](https://colab.research.google.com/drive/1h4yoaIYcDRHy6XYDtMgrUUCB0fJ1uf0v?usp=sharing). I'd like to offer this up as an...
Hi all, the [MatrixNormal](https://en.wikipedia.org/wiki/Matrix_normal_distribution) provides a clean way to parameterize the distribution of `n x k` matrices `X ~ MN_{nk}(M, U, V)` as a function of its `n x k`...
Hey all, would be nice if there was an implementation of [Automatic structured variational inference](https://arxiv.org/abs/2002.00643) in NumPyro (unless it is already there and I am not seeing it). A TFP...
Dear Numpyro developers, Please develop Euler Maruyama features in numpyro similar to features found in PyMC. Thanks alot.
Originally, this is used to test for funsor PR: https://github.com/pyro-ppl/funsor/pull/392. But I found that we need more tests when updating the code to match Pyro version. To pass some tests,...
Hello -- New to numpyro, but really excited to be looking at it! Question: When I call something like ``` nuts = NUTS(model) mcmc = MCMC(nuts, num_warmup=500, num_samples=2000) mcmc.run(rng_key_, data)...
Continue of #977. Pair-coded with @eb8680 @fritzo @ordabayevy ### TODO - [x] fix substituting issues of `enum_scan`: under `enum`, latent variable names will be `_PREV_z` and `z` while without it,...
Despite that NumPyro is very fast (comparing to other frameworks), running MCMC for large datasets is still slow. Even that GPU helps us increase 10x the speed, it is still...
This proposes to implement a vectorized interpretation of `pyro.scan` that completely parallelizes over the time axis. This follows a hand-implementation of vectorization in `pyro.contrib.epidemiology` (in [CompartmentalModel._relaxed_model()](https://github.com/pyro-ppl/pyro/blob/776b6ef966a089bce573cdf34a202a45f34053e2/pyro/contrib/epidemiology/compartmental.py#L980-L983) and [._vectorized_model](https://github.com/pyro-ppl/pyro/blob/776b6ef966a089bce573cdf34a202a45f34053e2/pyro/contrib/epidemiology/compartmental.py#L922-L946). This interpretation...
