We're going with the few core model designs / pipelines / spines approach to example/test management as opposed to adding every region
With that being said, I'll add a paragraph about an optional plan for testing agency contributed models as well
It would be optional and paid for separately
It would be coordinated / run by the team
It may be run nightly/weekly instead of with each commit
It would require private (paid for) data management and computing solutions on top of the existing free open source tools being used
I'll add a development roadmap to the contributions from others section since we want to be more explicit about the project's future
@PMC can review the updated strategic plan next week
Task 10 Support for Three Zone Systems and TVPB (milestone)
Jeff finished adaptive chunking, which will speed up existing models and help with TVPB memory management
Jeff now setting up the marin tm2 work tour mode choice example
After that, work on caching / parallelization
Should we do some more performance tuning (such as better manage data types) since Jeff is into that part of the code now? Since TVPB leads to some memory issues, should we maybe revise our approach? Let's discuss on Tuesday.