Steven G. Johnson
Steven G. Johnson
I was looking in [your build.jl file](https://github.com/malmaud/TensorFlow.jl/blob/e1de3839bef87e19a1543084c0c3dc1e93af5347/deps/build.jl), and it seems that [these lines](https://github.com/malmaud/TensorFlow.jl/blob/e1de3839bef87e19a1543084c0c3dc1e93af5347/deps/build.jl#L39-L59) could be considerably simplified to: ```jl pyimport_conda("tensorflow", "tensorflow=" * cur_py_version, "conda-forge") ``` This also has the advantage...
I notice that you have some custom code for parsing foreign version numbers via VersionNumber(string). You might try vparse from https://github.com/stevengj/VersionParsing.jl instead. ```jl julia> vparse("0.12.head") v"0.12.0+head" ```
It seems like this should work. All that is needed is: ```jl Base.big(x::Dual{T}) where {T} = Dual{T}(big(x.value), big(x.partials)) Base.big(p::Partials) = Partials(big.(p.values)) ``` (Mentioned [on discourse](https://discourse.julialang.org/t/optim-jl-cant-optimize-function-which-includes-bigfloat-variables/65095).)
> `nout`: The output size of the function f. Defaults to 1, i.e., a scalar integral output. It seems like this `nout` parameter should not be needed for all algorithms....
`datatype("this is a string")` works, but `datatype(String)` fails with a MethodError. Seems like a missing method, since e.g. both `datatype(1.0)` and `datatype(Float64)` work. Mentioned on [julia-users](https://groups.google.com/d/msg/julia-users/pUD-1psTUz8/jHyt45m8BAAJ).
Would be nice to have a docstring explaining the constructor arguments etcetera. ``` help?> JacVecOperator search: JacVecOperator AnalyticalJacVecOperator No documentation found. ``` Presumably there are other types that need documentation...
This PR adds a new function `openblas_set_threads_callback` that allows the caller to register a callback in order to change the threading backend at runtime — instead of spawning its own...
It looks like may be using `norm(matrix)`. In Julia 0.7, this will compute the Frobenius norm (`vecnorm` in Julia 0.6), due to JuliaLang/julia#27401. If you want the induced/operator norm as...
At some point, it might be worthwhile to move PythonCall.jl and friends to the JuliaPy organization. (Up to you, of course!)
It would be nice if the node size automatically stretched out to fit the label text. (Also, currently you can only change the node radius, but for this you would...