Skip to content

decouple initial + model eval samples, re-package datasets
Compare
Choose a tag to compare
@mathematicalmichael mathematicalmichael released this 17 Mar 03:47
· 37 commits to main since this release
v0.0.15
dbb70b9
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

this release introduces a lot of breaking changes (not backwards compatible at all) in the interest of cleaning up syntax and supporting the ability to specify a prior/initial independent of a sampling distribution used to evaluate samples. The concept of "sample_dist" is used to refer to the distribution which was used to evaluate the model. "dist" refers to the prior distribution. New args are "loc" and "scale" which are used for scipy's distributions to set the prior to one's choosing. "tolerance" used to refer to measurement tolerance (precision of sensor), and now has been replaced with "precision" or "-p" with "-t" and "tolerance" used exclusively to refer to the parameter which is used to define the standard deviation of the sampling distribution when it is set to be normal. Tolerance in this context refers to the support of the prior/initial. File format is pde_{dim}D/{1000*tol}_{dim}{dist}.pkl

Assets 2
Loading