Brandon T. Willard

> Updated [notebook](https://gist.github.com/xjing76/792a7bebfa954d176adb02d94913210b) > > 1 Increased number of obs to 500 2 Increased number of samples to 1000 3 Changed alpha prior to gamma distribution. 4 Added in a...

FYI: I've created some toy model tests in a PR [here](https://github.com/NCCMedia/survey-book-models/blob/e076730a89a2c0b22c4cfae7ea62bd05300a2ff3/src/org/survey-book-models.org#custom-samplers).

Here's a [Gist](https://gist.github.com/brandonwillard/b1b1a410d2e8f83761a7353c871a3373) that walks through the act of simulating and estimating the Poisson-Zero HMM model using the code in the `.org` file. In general, we need to - investigate...

Now that we have a Python package set up, the results of this work should be additions to the unit tests in `tests/`.

We're going to make an intermediate update that allows the use of both Aesara and the old Theano-PyMC. We can remove support for the latter once we transition to PyMC3...

Here are some references that cover (hierarchical) DP priors in HMMs: - ["Hierarchical Bayesian Nonparametric Models with Applications"](https://www.cs.princeton.edu/courses/archive/fall11/cos597C/reading/TehJor2010a.pdf), Chapter 3. - ["A Gentle Introduction to the Dirichlet Process, the Beta...

@stuartarchibald, this looks great! > * Is this enough for those interested to start work on implementing `numpy.random.Generator` more fully? If not what else would be useful? As far as...

Is there an issue that tracks the dispatch performance overhead problem (e.g. mentioned [here](https://github.com/numba/numba/pull/7900#discussion_r827973914))? While the underlying problem might not be specific to this functionality, it might still be good...

Sorry, I lost track of this, but is general `Generator` lowering covered in any of the active PRs, or is that a separate thing? For example, on 0.57.0dev0 I get...

On a related note, is there an "excluded"-like option for `vectorize` (i.e. similar to `numpy.vectorize`'s `excluded` option)? I'm trying to figure out how best to broadcast the scalar `Generator` methods...