batch size for posterior rejection sampling.

[janfb]

This will hopefully speed up rejection sampling from the posterior in case of leakage.

I moved the sampling and sample rejection to a separate function to reduce code repetition.

[michaeldeistler]

Thanks for tackling this! I saw too late that you had only commited and not yet requested review....so probably this is still WIP. Anyways, I had already written some comments, here they are ;)

[jan-matthis]

Just for future reference, a quick test showing the speed-up that can be obtained by batch sampling:

[codecov-commenter]

Codecov Report

Merging #340 into main will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #340      +/-   ##
==========================================
+ Coverage   73.62%   73.65%   +0.02%     
==========================================
  Files          48       48              
  Lines        2684     2687       +3     
==========================================
+ Hits         1976     1979       +3     
  Misses        708      708              

Flag	Coverage Δ
#unittests	73.65% <100.00%> (+0.02%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
sbi/utils/sbiutils.py	82.40% <100.00%> (+0.50%)	⬆️
sbi/inference/snre/snre_base.py	95.78% <0.00%> (ø)
sbi/inference/posteriors/base_posterior.py	66.83% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b758022...d05cd57. Read the comment docs.

[jan-matthis]

Looks good to me!