Releases: hammerlab/stancache
Releases · hammerlab/stancache
Redo release following merge error
Update requirements to include seaborn
bugfix
minor bugfix; update readme
Update config.py (#12) fix small error in restore_defaults
try again (push to pypi)
Previous deployment failed due to permissions error. Thus we updated the encrypted password & will try again.
First push to pypi
Stancache has utilities for saving pystan
fitted models to disk, such that the file name is unique given the stan model code & unique parameters passed in to fit the stan model. Stancache sets the default seed value for any model fit via config file, so that re-fitting models can be triggered by modifying the seed.
This avoids the expensive task of re-fitting the same model given identical inputs, and allows safe pickling/unpickling of model fit & compiled model code.