search

tapios / risk-networks

Code for risk networks: a blend of compartmental models, graphs, data assimilation and semi-supervised learning
Other
2 stars 2 forks source link

WIP: DA updates of the whole state vector #195

Open dburov190 opened 3 years ago

dburov190 commented 3 years ago

At the moment, this PR only implements full-state update for the no-batching case; more to come.

To use it, set --assimilation-update-record (and/or --assimilation-update-test) to full_global, and corresponding batching size to 1.