ForceBru

fix: equality comparison where assignment was likely meant

2

As per #976, it looks like in this code in `IPNewton`: https://github.com/JuliaNLSolvers/Optim.jl/blob/adc5b277b3f915c25233b45f8f2dd61006815e63/src/multivariate/solvers/constrained/ipnewton/ipnewton.jl#L282-L288 ...the line `state.ev == equality_violation(constraints, state)` probably meant assignment, not an equality test: 1. The equality test would...

Interior-point Newton seems to be non-deterministic

7

# Code ```julia import Pkg Pkg.activate(temp=true, io=devnull) Pkg.add(name="Optim", version="1.6.0", io=devnull) Pkg.status() const AV = AbstractVector{T} where T using DelimitedFiles, Optim normal_pdf(x::Real, μ::Real, var::Real) = exp(-(x-μ)^2 / (2var)) / sqrt(2π *...

Equality test instead of assignment in `IPNewton`?

1

`state.ev == equality_violation(constraints, state)` in this snippet of `update_state!` is an equality test, but it looks like assignment was meant: https://github.com/JuliaNLSolvers/Optim.jl/blob/adc5b277b3f915c25233b45f8f2dd61006815e63/src/multivariate/solvers/constrained/ipnewton/ipnewton.jl#L282-L288 Looks like currently the algorithm only updates `state.ev` here:...

Default `GA` and `DE` don't seem to work in v0.11.1

6

# TL;DR `GA()` and `DE()` don't move away from the initial point, say that any initial point is the optimum and _report convergence_, even though the algorithm isn't anywhere near...

`create_group` probably should throw a custom exception

3

Right now there's a call to `error`: https://github.com/JuliaIO/HDF5.jl/blob/ef228d57be54d2082a6a7581895fd4becbfe3714/src/HDF5.jl#L723 In my code, I'd like to overwrite the data in a group _if it already exists_, but currently that would make me...

`UndefRefError` when running `Enzyme.gradient!`

5

# Code ```julia import Pkg; Pkg.status() import Random import Enzyme const AV = AbstractVector{T} where T # ===== Set up objective function ===== normal_pdf(x::Real, mean::Real, var::Real) = exp(-(x - mean)^2...

conversion to pointer not defined for `SubArray{Float64, 1, Vector{Float64}, Tuple{UnitRange{UInt64}}, true}`

1

Below is a copy of the [issue](https://github.com/JuliaStats/Clustering.jl/issues/219) I opened in the Clustering.jl repo. In the comments of that issue it was [suggested](https://github.com/JuliaStats/Clustering.jl/issues/219#issuecomment-892885187) that I open an issue here, so here...

Can't sample from `MultivariateNormal` when covariance matrix is supposedly not positive definite, but NumPy can

2

__TL;DR__: given the same mean vector and covariance matrix, NumPy can sample from a multivariate normal with these parameters, but `Distributions.MultivariateNormal` says: `PosDefException: matrix is not positive definite; Cholesky factorization...

`pdf` is ambiguous for `UnivariateGMM`

1

## Code ```julia julia> distr = UnivariateGMM([-1, 1], [.1, .2], Categorical([.4, .6])) UnivariateGMM{Vector{Int64}, Vector{Float64}, Categorical{Float64, Vector{Float64}}}( K: 2 means: [-1, 1] stds: [0.1, 0.2] prior: Categorical{Float64, Vector{Float64}}(support=Base.OneTo(2), p=[0.4, 0.6]) )...

The only explanation of `filldist` I could find is here: https://turing.ml/dev/docs/using-turing/performancetips#special-care-for-tracker-and-zygote > `filldist(dist, N)` creates a multivariate distribution that is composed of `N` identical and independent copies of the univariate...