Frames White

sourcewalk introduces blocks

3

### MWE: ``` using MacroTools using MacroTools: textwalk code = """ Rule(x -> -sin(x)) Rule(x -> 1 + tan(x)^2) """; after = textwalk(code) do expr @capture(expr, Rule(v_)) && return v...

This is hard to explain. here is a demo I want to remove the second argument of a 3 argument function. ``` julia> using MacroTools: textwalk julia> code = """...

std(::AbstractVector{Float32}, ::AnalyticWeights{Float32}) gives a Float64

2

Setup: ```julia julia> using StatsBase, Statistics julia> xs = Float32[1, 2, 3, 4, 5]; julia> ws = AnalyticWeights(Float32[1, 1, 1, 1, 1]); julia> eltype(ws) Float32 ``` Now trying some operations...

For some reason v0.12.0 never actually got registered. it is needed to fix #515 - [ ] Correct UUID to match register - [ ] Add compat bounds for all...

Allow failure on Python 2, not 3

5

Since Python 3 is now the default https://github.com/JuliaPy/Conda.jl/pull/108

Add docs for Eager mode

1

Readme, re: eagermode is out of date

Broadcasting turns sideways, when dimensions have to be added

1

I'ld been wondering why I could never write networks with vector biases. Always had to use 2D column matrix basises. And now I understand. And it should have been obvious,...

I believe it is `@tfimport OperationName` But it is not in the docs.

seeL ``` using TensorFlow sess=Session(Graph()) bizz= Tensor(4) + Tensor(4) + Tensor(4) @tf begin foo = (Tensor(4) + Tensor(4)) + Tensor(4) bar = Tensor(4) + Tensor(4) + Tensor(4) end ``` **MethodError:**...

I am doing the thing where I write code that works for normal julia (when on AbstractArrays), and for TensorFlow (when on Tensors). Not being able to use `@view` is...