Tor Erlend Fjelde

Just to add to this issue: as @tpanum I tried following the advice of @arthurcgusmao and, also, still ran into issues. Eventually I managed to figure out that for, some...

I would also _really_ like to able to save any output to a file, e.g. if the results are HTML but I've specified `:file tmp.html` then the result should be...

One hacky way I got the HTML results to respect `:file` was to do: ```emacs-lisp (cl-defmethod jupyter-org-result ((_mime (eql :text/html)) content params) (if (alist-get :file params) (let ((fpath (alist-get :file...

> @pxl-th could you please also running your benchmarks on 0.6.29 and 0.6.28 to see if this is a recent regression, or a problem that has been around for a...

> @torfjelde for my case (NN using Flux.jl), compile times can be significantly improved by fixing type-stability issues with layers. See [FluxML/NNlib.jl#370](https://github.com/FluxML/NNlib.jl/pull/370) for timings, mainly first forward pass. But it...

Also one thing I'm noticing in our issue: the memory usage on some of the older versions, e.g. [email protected] is _insane_. This is _much_ better on more recent versions though.

> @torfjelde have you tried checking with jet.jl? Nope, but I just tried it but I have to use `[email protected]` because Turing doesn't support Julia 1.7 quite yet. Using `@analyze_call`...

To address the particular error from above, I believe the following should do it: ```julia using ChainRulesCore function ChainRulesCore.rrule(::typeof(getproperty), A::LArray, s::Symbol) function getproperty_LArray_adjoint(d) # NOTE: I hope this reference to...

Sure can :+1:

Bump:) What is needed for this to make it into CUDA.jl? Is it just a matter of moving the code to that package or is there anything else required?