nutpie

nutpie copied to clipboard

Right now everything in nutpie (the sampler and the logp functions) are working in float64. For some models we could get decent speedups if we also support float32. Pytensor can...

aseyboldt

If a deterministic variable uses a RandomVariable, it won't properly update the seed in the `expanding_function` of the compiled pymc model, as we only use the numba function but do...

aseyboldt

I figured out that you have to prefix the model name followed by two colons when updating variables in the `with_data` keyword arguments. I don't know if this is intended,...

a2435191

Allow pytorch backend when available in PyMC

2

The barebones PyTorch backend should be in in the next PyTensor release and subsequent PyMC. I think the best way to test it out and gather info on what Ops...

ricardoV94

Slow initialization running PyMC model

4

I am trying to run a PyMC model with the nutpie sampler, but initialization seems to be very slow and sampling does not start with these Warning statements: ``` C:\Users\TsubasaTakaya\miniconda3\envs\pymc_env\Lib\site-packages\pytensor\link\numba\dispatch\basic.py:379:...

ghost

Test compiling stan model in conda with conda-forge compilers installed

7

Otherwise it seems this can lead to the stan makefile not recognizing the compiler: ``` RuntimeError: Command make STAN_THREADS=true STANCFLAGS=--include-paths=. /tmp/tmp9av6ksrm/model_model.so failed with code 2. stdout: stderr: stan/lib/stan_math/make/libraries:117: *** "Need...

aseyboldt

updates: - [github.com/astral-sh/ruff-pre-commit: v0.5.6 → v0.6.7](https://github.com/astral-sh/ruff-pre-commit/compare/v0.5.6...v0.6.7)

pre-commit-ci[bot]

Is there a way to disable the jittering of the initial point on each chain when sampling? I'm working on quite fussy models, and would like finer control over the...

jessegrabowski

Document/expose from_pyfunc?

2

Hi! I think it is fairly common for a user to have a log density/its gradient _outside_ of a PyMC or Stan model. I was helping one such [user on...

WardBrian

WIP pytorch/onnx backend

1

This exports a torch module (or function) using dynamo_export to onnx, and then uses the onnxruntime to sample from the model. Still WIP, not tested much and the interface is...

aseyboldt