Here are some fixes I needed to do on Julia 1.0 to get the code running when models don't converge. Specifically,
Changed warn() (0.6 syntax) to @warn().
For glmnetcv, fits is a vector of 1 by n vectors, which potentially contains empty vectors. I filtered out empty vectors before concatenating them, otherwise a dimension mismatch error will be thrown.
Here are some fixes I needed to do on Julia 1.0 to get the code running when models don't converge. Specifically,
warn()
(0.6 syntax) to@warn()
.glmnetcv
,fits
is a vector of 1 by n vectors, which potentially contains empty vectors. I filtered out empty vectors before concatenating them, otherwise a dimension mismatch error will be thrown.