Carlos Parada
Carlos Parada
The changes to `eachslice` in 1.9 might cause trouble for this package. I'm working with DimensionalData at the moment, and there's problems with it at the moment--nothing breaks, but `eachslice`...
Spherical coordinates can be generalized to hyperspherical coordinates, as discussed [here](https://en.wikipedia.org/wiki/N-sphere#Spherical_coordinates).
I occassionally get extremely long/redundant stacktraces where the same error occurs many times in some part of the model. For example, the attached file: [stacktrace.txt.zip](https://github.com/JuliaAI/MLJBase.jl/files/13035623/stacktrace.txt.zip) Could this be improved somehow?
Most notably, several keywords (`unit`, `origin`) are completely undocumented. I also think that `scale` could be improved by using an ADT, like you'll find in MLStyle.jl; this has the advantages...
Could be useful for something like `DictTables`--it'd be super useful if I could index the `i`th row of a `DictTable` with an integer, or instead index it with a symbol....
The weighted method for `ash` seems to be undocumented. (@itsdebartha)
From the previous issue on calculations for `pointwise_loglikelihood`: > What do you expect `pointwise_loglikelihoods` to compute here? The marginal likelihood for each dimension? If so, I don't think this is...
I’ve noticed there’s no way to tell the difference between interval and ratio scales ATM. They’re both `Continuous` right now, but they’re not quite the same; ratio scales have a...
Declaring ranges in vectors creates type instabilities (which show up as false positives on JET.jl), besides feeling a bit weird and encouraging bad habits (using vectors for a small collection...
Now that we have [Optimization.jl](https://docs.sciml.ai/Optimization/stable/), would it be possible to support a broad class of optimizers for MLJTuning by just supporting the Optimization.jl interface?