Labels

This impacts the public API of the project (e.g. inference class).

Internal structure and design of the codebase

Good for newcomers

Something is in the way of fixing this. Refer to it in the issue

Something isn't working

Pull requests that update a dependency file

SBC, TARP, L-C2ST, coverage, sensitivity analysis, and plots

Improvements or additions to documentation

New default embedding nets

Improvement planned or wanted by the maintainers

Neural density estimators and the neural_nets module

Requested by a user, not yet planned by the maintainers

Pull requests that update GitHub Actions code

Device handling: CPU, CUDA, and MPS

For our in-person hackathons with maintainers present, not a general up-for-grabs list

Extra attention is needed

Trainers and potentials in sbi/inference

Installation, packaging, dependencies, and environments

Karma issues: potentially annoying but very important issues

Tracks failures of the latest-deps canary workflow

When the simulator does not match the observed data

Waiting for more information from the reporter

Posterior classes and their sampling interface

Pull requests that update python:uv code

Further information is requested

Needs experiments or method knowledge, not only an engineering fix

MCMC, VI, and rejection samplers

Score matching and flow matching methods, for example NPSE and FMPE

Everything to do with the sbi test suite

This will not be worked on