azev77

Domain Errors and Log Transforms

1

It works for the following set of parameters: ```julia using DifferentialEquations, Plots; function ngm!(dX, X, θ, t) C, K = X ρ, γ, z, α, δ = θ dC =...

Unexpected loading behavior

27

I installed a fresh version of Julia 1.6 on my macbook. 1 installed the latest MLJ 1.6 then I tried to auto load all MLJ relevant packages ```julia @inline function...

Model table like other interfaces: MLR3, Caret etc

4

I was comparing the MLJ [table of models](https://alan-turing-institute.github.io/MLJ.jl/dev/list_of_supported_models/) with: [MLR3](https://mlr3extralearners.mlr-org.com/articles/learners/list_learners.html) [Caret](https://topepo.github.io/caret/available-models.html) I found those tables have some nice searchable features. E.g. might it be worth adding a column for model...

Wrap a simple R model

1

Hey guys, I'd like to wrap [sprintr](https://cran.r-project.org/web/packages/sprintr/index.html). Can you point me to an example of wrapping an R model in MLJ? Here is the code for `fit` and `predict` ```julia...

Issues with some regression models [tbc]

14

```julia using MLJ, RDatasets, DataFrames; #Load & make model list. @inline function load_m(model_list) @inbounds for (i, model) in enumerate(model_list) load(model.name, pkg=model.package_name, verbosity=0) # end end # Load all models ONCE!...

models(matching(X, y)) for Images

2

Currently: `models(matching(X, y))` doesn't return relevant models when X has images ```julia import Flux; X, y = Flux.Data.MNIST.images(), Flux.Data.MNIST.labels() typeof(X), typeof(y) models(matching(X, y)) ``` I'm not sure if this is...

Proposed model

2

Suppose I got a new dataset in the mail today & wanna see which brand-name distribution in [Distributions.jl](https://github.com/JuliaStats/Distributions.jl/) best fits it. ```julia using Distributions, Random, HypothesisTests; Uni = subtypes(UnivariateDistribution) #Cts_Uni...

Infinite Dimensional Hyperparameters

1

Some models allow infinite dimensional hyper-parameters. **Eg 1**: [XGBoost](https://juliacomputing.com/blog/2020/02/24/ad-xgboost.html) allows custom loss functions. (As does @xiaodaigh's [JLBoost.jl](https://github.com/xiaodaigh/JLBoost.jl)) **Eg 2**: @joshday's [SparseRegression.jl](https://github.com/joshday/SparseRegression.jl) allows custom loss & custom penalty ```Julia s =...

Deadlinks

4

I submitted a few PRs w/ typos. Note, most of the tutorial pages have an "Edit me" button with dead links. [Introduction to Turing](https://turing.ml/dev/tutorials/00-introduction/) gives the following dead link: https://github.com/TuringLang/Turing.jl/edit/master/docs/src_tutorials/00_introduction.md...

Feature request: monotonicity constraints

2

Often when economists use neural networks, they have theoretical restrictions on the shape: Monotonically increasing Concave etc is this possible in Flux?