Anthony Blaom, PhD

# Proposed behaviour ## When should MLJ actions trigger mlflow logging? It should be possible to request logging for these actions: 1. **Serializing machines:** Calling `MLJModelInterface.save(location, mach)` whenever `location` is...

# Some miscellaneous thoughts on implementation 1. A [proof of concept](https://github.com/JuliaAI/MLJBase.jl/pull/912) already exists for performance evaluation. This shows how to add the new functionality using an extension module, which also...

cc @pebeto @deyandyankov @tlienart @darenasc

I don't know anything about DagsHub but I wonder if it doesn't just work already, as we already have [some integration](https://github.com/JuliaAI/MLJFlow.jl#mljflowjl) of MLJ with MLflow. @pebeto may be better able...

While I agree the suggested behaviour would be an improvement, I'm not sure I'd characterise this as a bug. Also, it's actually quite difficult to implement because there is a...

> As I mentioned in slack, I would like to help improve the combination of MLJ and DataFrames, as I think together they have potential for some really elegant workflows...

Thanks for raising this issue. `unpack` is designed to work with any table `X` satisfying `Tables.istable(X) == true` as it would for any `DataFrame`, so perhaps this is a bug....

Closing as stale (no further response from poster).

I think it would be good to add an example along those lines. Have opened [this](https://github.com/alan-turing-institute/MLJ.jl/issues/1055). A Wolpert stack trains an array of models on different parts of the training...

Correct. MLJ does not directly provide such a transformer. Ideally, this would be more than a `Static` transformer: You want to learn the redundant features by just looking at training...