mauro3 / kissmcmc.jl Goto Github PK

View Code? Open in Web Editor NEW

14.0 4.0 3.0 4.95 MB

Keep it simple, stupid, MCMC

License: Other

Julia 100.00%

mcmc-sampler emcee

kissmcmc.jl's People

Contributors

Stargazers

Watchers

Forkers

sefffal standardgalactic szabo137

kissmcmc.jl's Issues

Crash with BigFloats

I have a crash which I've narrowed down to the following... no idea whether it's this package's fault though:

using LinearAlgebra, KissMCMC

f(θs, ts) = [-t*exp(θ-exp(θ)*t)/length(θs) for t in ts, θ in θs]

g(v, ts) = begin
    j = f(v, ts)
    x = j' * j
    # s, ld = LinearAlgebra.logabsdet(x)
    # ifelse(s>0, Float64(ld), -Inf)
    d = det(x)
    d>0 ? Float64(log(d)) : -Inf
end

KissMCMC.emcee(v -> g(v, 1:3), [rand(3) for _ in 1:10]; niter=10^4) # fine!

versioninfo()
KissMCMC.emcee(v -> g(v, big.(1:3.0)), [big.(rand(3)) for _ in 1:10]; niter=10^4)

julia> versioninfo()
Julia Version 1.5.0
Commit 96786e22cc (2020-08-01 23:44 UTC)
Platform Info:
  OS: macOS (x86_64-apple-darwin18.7.0)
  CPU: Intel(R) Core(TM) i5-7360U CPU @ 2.30GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-9.0.1 (ORCJIT, skylake)
Environment:
  JULIA_NUM_THREADS = 4
  JULIA_PKG_SERVER = https://geo.pkg.julialang.org

julia> KissMCMC.emcee(v -> g(v, big.(1:3.0)), [big.(rand(3)) for _ in 1:10]; niter=10^4)
julia(303,0x70000e9e8000) malloc: *** error for object 0x7fb6d3ea8d70: pointer being freed was not allocated
julia(303,0x108ff35c0) malloc: *** error for object 0x7fb6d306b460: pointer being freed was not allocated
julia(303,0x70000e9e8000) malloc: *** set a breakpoint in malloc_error_break to debug
julia(303,0x108ff35c0) malloc: *** set a breakpoint in malloc_error_break to debug

signal (6): Abort trap: 6
in expression starting at REPL[5]:1
in expression starting at Rin expression starting at REPL[5]:1
__pthread_kill at /usr/lib/system/libsystem_kernel.dylib (unknown line)
Allocations: 34375251 (Pool: 34366235; Big: 9016); GC: 38
Abort trap: 6

(And the same on Julia 1.4.2)

Use threading

Allow passing in of additional argument to logpdf

For efficient in-place algorithm, often work-arrays (etc.) are needed. And for each chain a separate one is needed. Thus update signature logpdf(x) to logpdf(x, work).

TagBot trigger issue

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

Test sometimes fail

It seems the test:

metropolis: Test Failed at /home/pkgeval/.julia/packages/KissMCMC/z6iaT/test/runtests.jl:37
  Expression: all(abs.(std(thetas) - std_) .< abs.(std_ * tol))
Stacktrace:
 [1] test_mean_std(::Array{Array{Float64,1},1}, ::ATest, ::Float64) at /home/pkgeval/.julia/packages/KissMCMC/z6iaT/test/runtests.jl:37
 [2] macro expansion at /home/pkgeval/.julia/packages/KissMCMC/z6iaT/test/metro.jl:16 [inlined]
 [3] macro expansion at /workspace/srcdir/usr/share/julia/stdlib/v1.5/Test/src/Test.jl:1116 [inlined]
 [4] top-level scope at /home/pkgeval/.julia/packages/KissMCMC/z6iaT/test/metro.jl:3
Test Summary: | Pass  Fail  Total
metropolis    |   27     1     28

sometimes randomly fail (perhaps the tolerance is too tight or something).

mauro3 / kissmcmc.jl Goto Github PK

kissmcmc.jl's People

Contributors

Stargazers

Watchers

Forkers

kissmcmc.jl's Issues

Crash with BigFloats

Use threading

Allow passing in of additional argument to logpdf

TagBot trigger issue

Test sometimes fail

Image storage

Multi-threaded tests?

Reduce blobs

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent