xrtm-train
xrtm-train is the optimization and replay layer of the XRTM ecosystem.
Page role
This page is a package-map summary. Training workflows, replay harnesses, and
optimization code live in xrtm-org/train.
Use this package when
- you are doing replay, calibration, or backtesting work beyond the default product shell
- you want to test candidate changes against historical or controlled evaluation loops
- you need the deeper improvement path after learning the released compare/export surface
What it gives you
- Backtesting: stress-test strategies against historical snapshots
- Optimization: training/calibration loops that can actually change system behavior
- Simulation: large-scale experiment workflows above the lower package layers
Canonical reference
Ownership boundary
Package-level optimization and replay guidance belongs in train. Use
xrtm.org to explain where that work sits relative to the released product path.