Examples to reproduce RoME/IIF Solver freezing up.
Instantiate the environment described in julia/Project.toml and julia/Manifest.toml.
Take a look at the multimodal_calculations.bash
and outlier_calculations.bash
scripts. Feel free to vary the number of threads and processes used in there.
The scripts should be run from this repositorys root (otherwise relative file paths will not work I think).
Any output is written to a folder in the dataset directory. If it runs well I'd be happy to get those :)
This is the state of my code before any changes were made to it due to to RoME #418 . Feel free to change the julia/mmiSAM/evaluation/solve2DIncremental.jl to turn of any multithreading features or so.
Also feel free to play with the nullhypo, spreadNH or inflation parameter to get a better solution if it runs well. Thats what I intended to do before it freezed everytime.
Also some data in the outliers folder is not used by the bash-scripts yet because I could not get a plausible solution from it, feel free to try this and get something going with different parameters.
The --tukey parameter does not change anything, its just for logpath consistency with GTSAM ;)