dswah / sgcrfpy

SGCRFpy: Sparse Gaussian Conditional Random Fields in Python

dswah/sgcrfpy Issues

add graph visualization
Updated 8 years ago1
Port matrix operations to Cython for increased speed.
Updated 8 years ago
loss increases sometimes
Updated 8 years ago1
compare performance between covariance and correlation matrices
Updated 8 years ago1
how to do warm starts?
Updated 8 years ago
tune beta for line search
Updated 8 years ago2
should we have effect factorization in the package?
Updated 8 years ago
add max number of backtracking iterations, and early stopping
Updated 8 years ago2
infinite loss
Updated 8 years ago2
check bactracking line search
Closed 8 years ago10
if we continue fitting loss goes up
Closed 8 years ago4
check regularization is non-negative
Updated 8 years ago
sparsify lambda and theta and U and V
Updated 8 years ago2
illustrate types of graphs possible to model
Updated 8 years ago1
fix references formatting
Closed 8 years ago1
should we have n_iter as argument to `fit` ?
Updated 8 years ago
remove all Gammas
Updated 8 years ago3
replace all lambda inversions with cholesky versions
Closed 8 years ago2
fit should allow fitting from current estimate
Closed 8 years ago
need to do row-column updates in coordinate descent!!
Closed 8 years ago
don't need the __init__.py in a flat repo
Closed 8 years ago2