You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Because fitting a collection of models is important for hyper-parameter tuning (in particular lasso strength coeff $\lambda$), we should offer class that can store multiple fits, and can produce plots (altair for tracing certain muts?) like the shrinkage and sparsity, loss, and potentially correlation.
Note: a bunch of useful code for merging parameters and plotting exist in the spike analysis.
In addition to the dms vep pipeline I'm imagining a more standalone pipeline for multidms, so hopefully we can parallel-ize these fits using those. @sminot might have a great platform for hosting both such pipelines
The text was updated successfully, but these errors were encountered:
@jbloom related to our conversation last Monday.
Because fitting a collection of models is important for hyper-parameter tuning (in particular lasso strength coeff$\lambda$ ), we should offer class that can store multiple fits, and can produce plots (altair for tracing certain muts?) like the shrinkage and sparsity, loss, and potentially correlation.
Note: a bunch of useful code for merging parameters and plotting exist in the spike analysis.
In addition to the dms vep pipeline I'm imagining a more standalone pipeline for
multidms
, so hopefully we can parallel-ize these fits using those. @sminot might have a great platform for hosting both such pipelinesThe text was updated successfully, but these errors were encountered: