Closed tlienart closed 3 years ago
there's a problem with MLJBase (?) see https://github.com/alan-turing-institute/MLJBase.jl/issues/425
@tlienart have you tried deleting the MLJBase folder from your path_to_julia_folder/.julia/packages/
directory and the reinstalling MLJBase.
I'm unable to reproduce this issue on my system. Am running ubuntu by the way.
thanks Sam, yes, I did; seems like it might be a thing with JLSO or Memento ...
yeah. That's makes much sense. It seems MLJLinearModels
call MLJ.save
there may be an issue with MLJLinearModels but if you have a look at the issue on MLJBase the error happens independently of it
I still can't reproduce on my pc. @ablaom could you see if you can reproduce this on your mac
No I cannot reproduce. All good in a fresh env on my mac:
julia> versioninfo()
Julia Version 1.5.1
Commit 697e782ab8 (2020-08-25 20:08 UTC)
Platform Info:
OS: macOS (x86_64-apple-darwin19.5.0)
CPU: Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz
WORD_SIZE: 64
LIBM: libopenlibm
LLVM: libLLVM-9.0.1 (ORCJIT, skylake)
Environment:
JULIA_PATH = /Applications/Julia-1.4.app/Contents/Resources/julia/bin/julia
JULIA_NUM_THREADS = 5
The fail on 1.0 seems unrelated. It suggests that travis is not building proper R installation.
@tlienart I'm guessing you turned testing with R back on just for a one time test?
I heard that the latest release of Optim has some problems on Julia 1.0.
@tlienart I think this has been fixed (Optim >=1.0 require julia >=1.2 to work). Tests for julia v1.0 now pass. see logs. julia nightly tests fails. I suggest allowing failures for julia nightlies.
thanks sam, I updated the travis yml we'll see
lgtm
closes #79