Kai Xu

The new TFBoard also supports a plane called hyperparameters (https://www.tensorflow.org/tensorboard/r2/hyperparameter_tuning_with_hparams). How should one add this?

I'm very new to protocol buffer. For 1. I found the folder for hyperparameters is https://github.com/tensorflow/tensorboard/tree/master/tensorboard/plugins/hparams. Which files should I try to convert? And how do one use `ProtoBuf.jl`? I...

I also managed to make it work using the code patch (which I made avaiable at https://github.com/xukai92/isync.git). I simply modified the formula of isync in my Homebrew (i.e. removing `url,sha256`...

@cpfiffer This maybe another target to test the new interface.

By looking at there interface, I think we can esaily wrap that library into Turing as it only relies on the likelihood of the model.

It looks like the signature `Gibbs(args::T...) where {T

> Pair{ struct A{T} v::T function A{T}(v::T) where {T} new{T}(v) end end julia> function A(args::T...) where {T A((1.0, 2), (3, 4)) ERROR: MethodError: no method matching A(::Tuple{Float64, Int64}, ::Tuple{Int64, Int64})...

It was a typo in my reply. I actually tried what you suggested: ```julia julia> struct A{T} v::T function A{T}(v::T) where {T} new{T}(v) end end julia> function A(args::T...) where {T

I added `N` to the Gibbs struct and its constuctors now.

There is a more comprehensive one avaiable here: https://github.com/TuringLang/TuringExamples/blob/kx/improve_functionality/benchmarks/auto_diff/loop-vs-unrolled.ipynb