Removed the optional correction of the precision matrix

[famura]

I suggest to remove the optional correction of the precision matrix for 2 reasons:

1. Forcefully making a matrix positive definite is hacky, and
2. This option was only useful for the evil case where SNPE-A yields a denisty esimator with lower precision than the proposal prior. Hoewever, if this is the case, rejection sampling from MoG typically "fails" (super high rejection rates). Since MCMC sampling is disabled for SNPE-A (makes sense), I think it is best to let the attept fail and signal the user to try different params or a different random seed 🙃

[codecov-commenter]

Codecov Report

Merging #497 (907bf46) into main (3c3a79e) will increase coverage by 0.00%.
The diff coverage is 40.00%.

@@           Coverage Diff           @@
##             main     #497   +/-   ##
=======================================
  Coverage   67.69%   67.70%           
=======================================
  Files          55       55           
  Lines        3969     3967    -2     
=======================================
- Hits         2687     2686    -1     
+ Misses       1282     1281    -1     

Flag	Coverage Δ
unittests	67.70% <40.00%> (+<0.01%)	⬆️

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

Impacted Files	Coverage Δ
sbi/inference/snpe/snpe_a.py	66.50% <40.00%> (+0.15%)	⬆️

---

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 3c3a79e...907bf46. Read the comment docs.