search

dfdx / XGrad.jl

eXpression gradients in Julia
Other
3 stars 4 forks source link

UndefVarError: autoselect_codegen not defined #10

Closed cscherrer closed 6 years ago

cscherrer commented 6 years ago

Following the tutorial:

julia> versioninfo()
Julia Version 0.6.0
Commit 9036443 (2017-06-19 13:05 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
  WORD_SIZE: 64
  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Prescott)
  LAPACK: libopenblas64_
  LIBM: libopenlibm
  LLVM: libLLVM-3.9.1 (ORCJIT, broadwell)

julia> Pkg.clone("https://github.com/dfdx/XGrad.jl")
INFO: Cloning XGrad from https://github.com/dfdx/XGrad.jl
INFO: Computing changes...
INFO: Installing Einsum v0.1.0
INFO: Installing Espresso v0.3.0
INFO: Installing Matcha v0.1.1
INFO: Installing Sugar v0.4.1

julia> using XGrad

julia> xdiff(:(y = sum(W * x .+ b)); W=rand(3,4), x = rand(4), b = rand(3))
ERROR: UndefVarError: autoselect_codegen not defined
Stacktrace:
 [1] macro expansion at /home/chad/.julia/v0.6/Espresso/src/utils.jl:57 [inlined]
 [2] #xdiff#29(::Dict{Any,Any}, ::Array{Any,1}, ::Function, ::Expr) at /home/chad/.julia/v0.6/XGrad/src/xdiff.jl:227
 [3] (::XGrad.#kw##xdiff)(::Array{Any,1}, ::XGrad.#xdiff, ::Expr) at ./<missing>:0
dfdx commented 6 years ago

Please also checkout latest version of Espresso.jl:

Pkg.checkout("Espresso")

XGrad.jl isn't registered yet, so it may depend on untagged versions of some packages.