View source
bnpy_soumya
  • Contributors
    1. Loading...
Author Commit Message Date Builds
250+ commits behind master.
sghosh
Various edits that work towards adding sampling capability
sghosh
Collapsed Gibbs Sampling support. Note the code doesn't yet play well with Run.py
sghosh
Added support for multiple likelihood mixture models. Works with EM and VB learning algorithms. Currently only works with multiple Gaussian Likelihoods FIXED: 1) BNPYArgParser now supports multiple observation models TO DO: 1) Fix ModelWriter 2) Add Multinomial support 3) Add admix support
sghosh
Batch variational inference algorithms now support multiple Gaussian observation models. 1) FIX: Clean up argument parsing. 2) TO DO: Add online inference functionality. 2) Test: Test multinomial observation models.
sghosh
Hacky version of EM with multiple likelihoods now works. + Merge branch 'master' of https://bitbucket.org/michaelchughes/bnpy-dev into bnpy_soumya # By Mike Hughes # Via Mike Hughes * 'master' of https://bitbucket.org/michaelchughes/bnpy-dev: FIX moVB now works when calling bnpy.Run.run from within python
sghosh
Minor edits
sghosh
Resolved conflicts merging files in bnpy/distr
sghosh
Changed sample function for ZMGaussDistr
sghosh
Minor edits: More informative doc strings.
sghosh
no message
sghosh
Various edits to sampling functions and added test cases.
sghosh
Can now sample from Wishart and GaussWishart distributions
sghosh
Added function for reading images.
sghosh
Added sampling support for basic distributions
sghosh
Created new branch bnpy_soumya