Anthony Blaom, PhD
Anthony Blaom, PhD
A related discussion: https://github.com/FluxML/MLJFlux.jl/issues/97 .
@lorenzoh
That sounds great. Hope this works out.
@darsnack @lorenzoh Am really appreciating this discussion! Will try to add some comments early next week. Thanks again.
Okay, I've had a chance to educate myself a little better about MLDataPattern and DataLoaders. It may make more sense to integrate MLJ with MLDataPattern first, after all. This will...
@ashryaagr CompatHelper it seems the ssh keys are not working. It's worthwhile resetting these, in my experience.
The other idea, mentioned on the call (which does not assume you have existing structs `MLJFair.TruePositive` and so forth) would be a wrapper. So user does something like `m =...
Completed.
The syntax `fit!(mach, X, y)` sounds like a good suggestion - we probably don't need to separately attach new data to the machine and then train. However, I can't see...
> So TL;DR my suggestion was that models would be required to implement something like fit!(model, X, y) Yeah, we have already have the stub (see [above comment](https://github.com/alan-turing-institute/MLJ.jl/issues/60#issuecomment-549589603)): ``` MLJModelInterface.online_update(model::Model,...