Guillaume Dalle

Results 1600 comments of Guillaume Dalle

My bad, I was wrong on the meaning of a training epoch! Thanks for your answer, I will try reducing batch size

Just bumping this to say that I would really love a `fit` interface in MeasureTheory that plays nice with Distributions, but also enables us to do more general things like...

I think it is feasible to add a few lines to `distproxies.jl` and extend the `fit` function in a quick and dirty way. The only challenge is that we need...

I'll get to work on that. Btw, do we really need both the `DistributionMeasure` type and our own rip-offs from Distributions.jl?

More generally, I think the docs contain a (Gaussian) mixture of LaTeX and Unicode symbols, so a global convention would be lovely

Bumping this to say that in my (ongoing) docs revamp, I'm trying to generally adhere to JuliaFormatter / BlueStyle, especially by grouping imports and exports

Following up on this, I like the idea of an interface with numerical integration. Could it be a small side package (to outsource the Quadrature dependencies) that mostly defines, say,...

Yes but in that case it means that you cannot apply a measure to arbitrary sets. For instance, if you have a Poisson process on R^n but you want to...

What kind of additional functions did you have in mind?

I think the `sortperm` is useful for quick `cdf` computations, because then you only need a `searchsortedfirst(vs, x)` to know at which step of the discrete distribution you are.