Jaime RZ

Hi Tor! Thanks for opening an issue. After some thought I am also in favour of 1; i.e. `Turing.sample` . I think besides the points you have raised I can...

I have been playing around with these ideas and I have noticed that even if we get rid of `InferenceAlgorithm` would still need to wrap AbstractSampler in DynamicPPL.Sampler() if we...

Quick note: If we do this we will have the same type ambiguity that we would have in `sample` in `bundle_samples`.

Hi All! My name is Jaime Ruiz. I am cosmology PhD in the Universiy of Oxford working with David Alonso and Pedro Ferreira. I also recently joined the Turing.jl org....

Ok that's interesting. I was proposing to get rid of both but keep the instance of `AbstractMCMC.sample` seems reasonable to me. One thing that I am doing in #325 is...

I think even though #325 is now merged. Having this function here is probably still useful if users want to use the AbstractMCMC interface without loading Turing. However, I would...