Anthony Blaom, PhD
Anthony Blaom, PhD
Comment from @ven-k on slack: --- While defining the struct for losses, including the y slightly improved the time taken. For ex, ``` struct MSE{T
**LossFunctions** fix: We can make measures from LossFunctions behave exactly like all the others when called by *importing* their names into scope, instead of `using` them, and exporting versions that...
I believe we already have negative log-likelihood, aka log-loss. It is called `cross_entropy` and, yes, it is a proper scoring loss. ```julia search: cross_entropy cross_entropy Cross entropy loss with probabilities...
Ah yes. `cross_entropy` is for `Finite` targets only. And yes, it has an extension to the continuous case (https://www-jstor-org.ezproxy.auckland.ac.nz/stable/2629907?seq=3&socuuid=dcad1753-575a-42c3-a7b8-fdd39a9f7589&socplat=email#metadata_info_tab_contents ) but it is not yet implemented. Indeed no measure for...
https://github.com/alan-turing-institute/MLJBase.jl/issues/450
Lighthouse has some measures we may want to include: https://github.com/JuliaAI/MLJBase.jl/issues/586
Community discussion on mitigating metric code fragmentation https://github.com/FluxML/ML-Coordination-Tracker/discussions/38
This is great to have this review. However, generally models take *far* longer to train than they do do evaluate, and whence performance optimisation here has never been a priority...
@DilumAluthge Can this be closed now?
I guess this sounds silly, but I don't like all those red crosses I get after my commits when nightly fails, which it frequently does. That's why I disabled nightly....