Penelope Yong comments

Results 323 comments of


                                            Penelope Yong

Adds `@returned_quantities` macro

Personally, I would be wary of introducing too many overloads. There is a question of whether the benefits of doing this (convenience, elegance, etc - these are more immediate) outweigh...

Non-dev docs not deploying

Fixed now by the addition of the key and deployment of a new package version :)

Implement `quantile(model(), u)`?

Hi @mvsoom - how would this work with multivariate distributions?

Implement `quantile(model(), u)`?

This is interesting. Sorry for the delay in response, I've been pondering it. It seems to me we would first need to introduce a QuantileBijector in Bijectors.jl, which obeys the...

Implement `quantile(model(), u)`?

Oh, as it happens, there's already an existing PR for the univariate one 😄 https://github.com/TuringLang/Bijectors.jl/pull/273 although what I called `QuantileBijector` in my post above is `CDFBijector` in this PR, and...

Add default implementations of `assume` and `dot_assume`

Probably no longer needed since `SampleFromPrior` etc. are being removed.

Remove `default_varinfo` function

@torfjelde, I'm sorry, you stumbled across a PRs which is definitely not intended for review yet 😅 I've been trying to see how much of the codebase I can strip...

Remove `default_varinfo` function

Eh, I can't remember what I was trying to do anymore. :// It was something with the experimental Gibbs sampler and trying to trace where things were going. Anyway, I...

Incorrect forward mode gradient for truncated normal distribution

```julia import Enzyme import FiniteDifferences using SpecialFunctions: erfc function g(i, _not_used) k = sin(sin(sin(1 * erfc(0.25) / 1))) return (i, k) end function f(_not_used) i = (0.0, 3.9555) t =...