Dehann Fourie
Dehann Fourie
- n-dimension using something like reshape or views somehow - add uniform per grid size over underlying AliasingScalarSampler - avoid the problem with AliasScSm where new samples are exactly at...
Should `variableType` not just exist at `DFGVariable` level? Why all the way down to VND...
MWE `useMsgLikelihoods=false` https://github.com/JuliaRobotics/RoME.jl/blob/f518061ea899434a52959e6bcc1a320626fdca76/test/testBeehiveGrow.jl#L28 ![Screenshot from 2021-05-17 08-44-50](https://user-images.githubusercontent.com/6412556/118490970-87b11a80-b6ec-11eb-80b5-1ca2b0c59aca.png) ---- - works better with intermediate solution `useMsgLikelihoods=true`, - related to #1010 - xref JuliaRobotics/RoME.jl#440
Ongoing and further consolidation of factors and different solvers. FYI @Affie
Solving with 10 processes, new RoME.beehive with 56 variables: ```julia signal (11): Segmentation fault in expression starting at REPL[17]:1 uv_fileno at /workspace/srcdir/libuv/src/unix/core.c:764 _fd at ./stream.jl:1124 [inlined] show at ./stream.jl:263 unknown...
Towards resolving #1010, and JuliaRobotics/RoME.jl#244
See potential breaking changes from JuliaLang/julia#40623 for Julia 1.7
MWE https://github.com/JuliaRobotics/Caesar.jl/blob/7167be93d3eef646ff6293e8398f8dfca7c8ef60/examples/dev/scalar/boxy.jl#L280 ```julia julia> pts = approxConv(fg, :x0x4f1, :x4) ERROR: DimensionMismatch("array could not be broadcast to match destination") Stacktrace: [1] check_broadcast_shape @ ./broadcast.jl:520 [inlined] [2] check_broadcast_axes @ ./broadcast.jl:523 [inlined] [3]...
We need to build a bridge between Manifolds and Optimization routines. This issue is just a place to list ongoing conversations that are already happening in the community. - JuliaManifolds/Manopt.jl#4...
Do a warning if there are no priors on a piece of a graph