From Slack:

Piping to papermill is on master 🤤
cat original.ipynb | papermill > result.ipynb

@simonbyrne noticed that the Gallery/Tutorial is slow to load. we should think about breaking it up into sub-pages, or using SVG instead of SVGJS, or ...

It's easy to implement single-site versions of these, where the user provides a (finite) set of values of the variable to consider. These would go into the inference library:

Map optimization:

function Gen.map_optimize(trace, addr, values)
    args = get_args(trace)
    argdiffs = map((_) -> NoChange(), args)
    weights = Vector{Float64}(undef, length(values))
    traces = Vecto

This issue will be used to track common interface option handling. This is a constant time factor so it's not the biggest deal, but I feel that we can improve some applications by reducing common latency here.

This is a follow-up on my question from the Gitter channel.
I am trying to create the following SDP with LMI constraints:

min c' x
s.t.  A0 + A1 * x[1] + A2 * x[2] >= 0

Assume that the matrices A0, A1 and A2 are sparse and the matrices A1 and A2 are stored in B, which is an Array{SparseMatrixCSC{Float64,Int64}}

using JuMP, LinearAlgebra, SparseArrays

m = 

The GR backend has implemented an :outer legend position, which is not reflected in the documentation.

Motivated by the following experience and other similar ones observed by colleagues, this issue is about improving the project page's usability for new users.

Watching this person use cocalc for the first time was a little disturbing...

• Clicked sign in with github. It worked! Didn't have to check the agree to terms button...
• To make a new project he tried to click new project button b

Readme, re: eagermode is out of date

I found that function mod2pi is not implemented yet, but mod works. Is there any list of implemented functions? Minimal working example is:

using Zygote

# This is working
gradient(x -> mod(x, 2pi), 1.)

# This is not
gradient(x -> mod2pi(x), 1.)

Here's a great opportunity for new contributors.

The main breaking changes effecting the examples would concern:

• tuning

• learning networks

These are now both covered in the updated manual, so that should be easy to sort. Note there are now two ways to export a learning network. The examples should use the new easy @from_netw

We still need docstrings for:

• Fminbox
• NewtonTrustRegion
• SimulatedAnnealing
• ParticleSwarm
• NelderMead

The other doc strings should be plenty inspiration if you're in doubt of what a doc string should contain for these types/constructors.

I was trying to use the Julia: Execute Code Cell Command, but I didn't know what was the delimiter for a cell. Finally, I found it, but It would be nice to have a comment about it :).

Add a section to the README that discusses the tiers of support, re. the recent Julia update to the README by @ViralBShah.

I tried to run Control example, but it does not work.

The error message said

MethodError: no method matching +(::Array{Any,1}, ::EqConstraint)
Closest candidates are:
+(::Any, ::Any, !Matched::Any, !Matched::Any...) at operators.jl:502
+(::Array, !Matched::Array...) at arraymath.jl:44
+(::Array, !Matched::SparseArrays.SparseMatrixCSC) at C:\cygwin\home\Administrator\buildbot\wor