Fix mnle tests, MCMC init pbar, sir batch size.

[janfb]

a couple of fixes for MNLE tests.

• test mnle with conditions only for 1D discrete
• fix vectorized evaluation of BinomialGammaPotential
• fix sir init function
• add pbar for sequential MCMC inits in general

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 78.18%. Comparing base (dd4aef7) to head (e6c56be).
Report is 2 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1410       +/-   ##
===========================================
- Coverage   89.20%   78.18%   -11.03%     
===========================================
  Files         119      119               
  Lines        8877     8916       +39     
===========================================
- Hits         7919     6971      -948     
- Misses        958     1945      +987     

Flag	Coverage Δ
unittests	78.18% <100.00%> (-11.03%)	⬇️

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

Files with missing lines	Coverage Δ
sbi/inference/posteriors/mcmc_posterior.py	83.58% <100.00%> (-2.57%)	⬇️
sbi/samplers/importance/sir.py	100.00% <100.00%> (ø)

... and 37 files with indirect coverage changes

[manuelgloeckler]

Thanks for fixing this!

The change in SIR is the bug fix, right?

Otherwise I suppose the most notebook changes are clean up to remove the warnings? If not I would remove them from the PR.