search

sbi-benchmark / sbibm

Simulation-based inference benchmark
https://sbi-benchmark.github.io
MIT License
88 stars 34 forks source link

Improve docstring for SL #38

Closed psteinb closed 2 years ago

psteinb commented 2 years ago

Looking forward to feedback on this. Haven't checked the sbibm paper yet, if it would lend a better explanation.

jan-matthis commented 2 years ago

Thanks for this!

Some extra context: We used SynthLikNet as a quick implementation of the classical synthetic likelihood algorithm published by Wood (2010). It is reported as SL on the website and in the paper's appendix.

psteinb commented 2 years ago

Cool, I'll have a look and add the reference in due course.

psteinb commented 2 years ago

Ready to go from my side @jan-matthis

psteinb commented 2 years ago

The test for sl.py I added fails at this point. But I gather that have a failing test is better than having no test.

jan-matthis commented 2 years ago

Cheers!