Closed joseortiz3 closed 7 months ago
I accidentally had the NormalFixedMean class parameterized by sigma/stdev (params[0] was the standard deviation). I changed it to match the Normal class's log(sigma).
@joseortiz3 can you pull the latest version of master and resolve conflicts here? Thanks and sorry for the delay in getting this merged
I implemented the distribution for NormalFixedMean (mean=0) closely following the code for NormalFixedVar (var=1) and Normal. Would appreciate someone double-checking the math for the d_score and metric. I also didn't see NormalFixedVar mentioned in the tests, so I added them to
test_distns.py
.Example plots: