Jan

SNLE and SNRE do not support Importance Sampling in build_posterior

1

## Description: A limitation in the current implementations of SNLE (Sequential Neural Likelihood Estimation) and SNRE (Sequential Neural Ratio Estimation), is that when attempting to use `ImportanceSamplingPosterior` through the `build_posterior`...

pairplot: documentation, more options and tutorial

2

The `pairplot` function is our main function for visualizing posteriors. https://github.com/sbi-dev/sbi/blob/c3c2b6e142fb4c57d5599effc1f01f8222a37c57/sbi/analysis/plot.py#L280 It would be great to improve it bit: - adding more documentation - defaults that produce a decent default...

## Problem The current setup for using Pyro MCMC methods is a bit of a mess: - we have a pyro-specific MCMC class in `sbi/samples/mcmc/mcmc.py` [1] - a pyro-specific implementation...

**Is your feature request related to a problem? Please describe.** For MCMC sampling we provide our own `slice_np_vectorized` MCMC samplers, and an interface to `Pyro` HMC and NUTS samplers. The...

Improve documentation website, add automatic API docs

1

Related to #120 More context: we build the docs by manually adding new classes or methods we want show on the website in the `API Reference` section here: https://github.com/sbi-dev/sbi/blob/main/docs/docs/reference.md and...

## Problem MNLE is a variant of NLE, tailored to cases where `x` is discrete and continuous. It has its down density estimator, `MixedDensityEstimator` living here: https://github.com/sbi-dev/sbi/blob/main/sbi/neural_nets/mnle.py It would be...

score-based diffusion models as density estimators

6

**Is your feature request related to a problem? Please describe.** Score estimation has been introduced for SBI, see - Compositional score estimation, good for iid data: https://proceedings.mlr.press/v202/geffner23a.html - sequential score...

tests are failing because of new pyright release.

3

there seem to be new checks in the pyright release that lead to pyright failing in our workflow. The latest commit on `main` was tested with pyright `v1.1.304`. the newest...

refactor: code repetition in train methods

1

## Description: The current implementation of the SBI library contains significant code duplication within the `train(...)` methods of SNPE, SNRE, and SNLE. These methods share many common functionalities, including: -...

ImportanceSamplingPosterior.sample is missing show_progress_bars argument.

2

## Description: Unlike other posterior classes in the SBI library, the `ImportanceSamplingPosterior` class `sample` method does not currently offer the option to enable or disable progress bars. This feature would...