byt143

14 ms·

91.
by byt143 8y ago
Still a very restricted subset of python, can't autodiff custom types etc
92.
by byt143 8y ago
Two of your issues are currently being worked on. The autodiff tracker stuff is temporary until the lower overhead and almost invisible compiler based AD mentioned in the blog post is fully ready. No custom types needed. There are also var
93.
by byt143 8y ago
Except much faster and much more generic. Eager execution is very limited in the subset of python it can differentiate, for example. There's more info in the OP blog post.
94.
by byt143 8y ago
I believe Julia has generic kernels for many linear algebra ops (like matrix multiply) that the compiler can use to generate code for different backends and types as a fallback to BLAS.
95.
by byt143 8y ago
Check out Makie https://github.com/JuliaPlots/Makie.jl And the cool thing is, Julia's wonderful generic programming facilities mentioned in the blogpost are used to make Makie generic to backend (GL, WebGL, cairo
96.
by byt143 8y ago
AD arbitrary julia code with mutation, closures, built in and user defined functions, fancy data structures etc (given the algorithm itself can be differentiated).
97.
by byt143 8y ago
Better support for static compilation Is on the roadmap
98.
Differentiable programming and deep learning with Julia(arxiv.org)
4 points by byt143 8y ago|0 comments
99.
by byt143 8y ago
Julia has it's own "*verse" type data pipeline framework, with an even greater variety of backends and plotting solutions than R. It's still in development (mutate and select have PRs) but it's almost there. https:
100.
Automatic Compilation of Julia Programs and Generic ML Models to Cloud TPUs(arxiv.org)
7 points by byt143 8y ago|2 comments
101.
by byt143 8y ago
Here is how to do explicit SIMD in julia: https://github.com/eschnett/SIMD.jl
102.
by byt143 8y ago
Actually Julia is aggressively inlined because it's easy to statically determine function calls at compile time...that's why the libraries are able to use such nice abstractions. It's already ahead of time compilable, just ne
103.
by byt143 8y ago
>Your code does tend to end up a bit c-like in those performance critical sections, but that's hardly surprising. That isn't true. Closures, higher order functions and fused broadcast array expressions are all very fast except
104.
by byt143 8y ago
Yes you can.
105.
by byt143 8y ago
Also that exact C example can be written almost verbatim in Julia, so C isn't "better suited" for even these small kernels, because one can drop to low level with even clearer syntax in julia
106.
by byt143 8y ago
https://scholar.google.com/scholar?cluster=26077808520337143...
107.
by byt143 8y ago
Something like this? https://github.com/andyferris/Traitor.jl
108.
by byt143 8y ago
Flux.jl is backend agnostic, so it has a pure Julia (with GPU code generation) and in progress ONNX and tensorflowjs export capabilities . And then once Julia can reliably compile to web assembly (already being worked on), your entire a
109.
by byt143 8y ago
It's going to have enough comparative advantage when a bit more mature that I can see it drawing off from the python community and entrenching itself as a serious contender.
110.
Travis Oliphant Leaving Anaconda, Inc(twitter.com)
4 points by byt143 9y ago|0 comments
111.
Popper as a particular case to Bayesian probability theory(soest.hawaii.edu)
2 points by byt143 9y ago|0 comments
112.
by byt143 9y ago
Julia is open source...why don't they just incorporate it?