Issues
- 2
- 0
- 1
add more type annotations to class args
#258 opened - 1
replace LGSSMParamsMoment with LGSSMParams
#257 opened - 1
- 0
Remove redundant GGF and iterated CMGF
#249 opened - 0
- 2
CategoricalHMM issues
#246 opened - 1
Standardize EM and SGD outputs
#241 opened - 1
- 0
Compare CMGF to memory-buffered SGD
#232 opened - 1
- 1
- 5
lgssm inference fails when run without bias.
#224 opened - 1
Failing MVN*HMM tests
#221 opened - 0
Add posterior_sample method to CMGF
#219 opened - 1
parallel LDS inference
#217 opened - 0
Implement Decoupled EKF for MLP Training
#215 opened - 0
Implement "fading memory" trick to CMGF
#211 opened - 0
Add a `LinearGaussianPoissonSSM` class
#210 opened - 0
- 0
Remove redundant Parameters containers
#204 opened - 0
Compare CMGF logreg to PASS-GLM
#199 opened - 0
Compare CMGF-MLP-classifier to one-pass SGD
#198 opened - 0
- 0
- 0
- 0
- 0
- 0
Create CMGF-MLP-Classifier Demo
#191 opened - 0
Create CMGF Poisson Likelihood Demo
#190 opened - 0
Separate the conjugate case from the current LinearGaussianSSM class and add learning algorithms
#181 opened - 0
make kmeans scaling plot
#180 opened - 1
Investigate the use of jax.tree_util.Partial for passing functions as arguments to jitted functions.
#179 opened - 0
- 0
References to batch_emissions.size don't work when you have JointDistribution emission objects
#173 opened - 1
fix m step in gaussian_hmm
#166 opened - 0
Implement logistic regression demo for CMGF
#158 opened - 1
- 0
Remove torch dependency
#152 opened - 1
Let's not copy/paste code from other repos
#151 opened - 1
- 5
refactor LGSMM class into subclasses
#149 opened - 2
replace distrax.* with tfd.*
#148 opened - 0
- 0
update GHMM M step to use NIW-update
#146 opened - 1
- 0
rename some directories / classes
#144 opened - 0
refactor fit_sgd to use TFDS instead of passing in entire data and creating our own minibatches
#143 opened - 0