cossio
cossio
Related: https://github.com/JuliaLang/Pkg.jl/issues/1982
We should use https://github.com/JuliaStats/LogExpFunctions.jl, which doesn't depend on Rmath. Note that StatsFuns just re-exports the functions from LogExpFunctions. See https://github.com/FluxML/NNlib.jl/issues/331.
I think I'm hitting a similar error, but on Linux. ``` julia> ENV["R_HOME"] "*" julia> using Pkg julia> Pkg.build("RCall") Building CodecZlib → `~/julia_depot_cossio/packages/CodecZlib/DAjXH/deps/build.log` Building Conda ────→ `~/julia_depot_cossio/packages/Conda/uQitS/deps/build.log` Building RCall ────→...
@simonbyrne I'm not sure it's the same issue so I separated it.
I would also like to have this. Any updates?
Does this look too small? $$\sum_{i=1}^N f_i$$ It does to me. Is there a way to make equations larger?
Is there anything holding back a 1.0 release at this point?
If I got this right, when you define a rule via `@scalar_rule`, it does *not* get applied when you broadcast that function?
What happens for `f: N×M -> PxQ`? Is the Jacobian `P x Q x N x M`?
Is this issue still an issue? There is no longer a `format` function.