# \[ANN\] Flux v0.8

**URL:** <https://discourse.julialang.org/t/ann-flux-v0-8/22305>\
**Category:** Machine Learning\
**Created:** [March 25, 2019, 3:15pm UTC](https://discourse.julialang.org/t/ann-flux-v0-8/22305 "2019-03-25T15:15:20Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![dhairyagandhi96](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/dhairyagandhi96/32/7589_2.png) [@dhairyagandhi96](https://discourse.julialang.org/u/dhairyagandhi96)\
**Post date:** [March 25, 2019, 3:15pm UTC](https://discourse.julialang.org/t/ann-flux-v0-8/22305/1 "2019-03-25T15:15:20Z")

</div>

We are pleased to announce that [Flux](https://github.com/FluxML/Flux.jl) v0.8 has been released and is available for download now. Here are some of the major highlights from this release and the full list is available in the [NEWS.md](https://github.com/FluxML/Flux.jl/blob/master/NEWS.md):

1. New [ConvTranspose layer](https://github.com/FluxML/Flux.jl/pull/311) @tejank10
2. Datasets are now [hash verified on download](https://github.com/FluxML/Flux.jl/pull/585) to avoid corruption. @KristofferC
3. [NamedTuples now work like Tuples](https://github.com/FluxML/Flux.jl/pull/603) when doing mapleaves @Keno
4. Added “[Performance Tips](https://fluxml.ai/Flux.jl/stable/performance/)” section to the documentation thanks to @oxinabox
5. The [training loop has been revamped](https://github.com/FluxML/Flux.jl/pull/651) and better shows the lower level APIs @oxinabox @MikeInnes
6. [Support for CuPtr](https://github.com/FluxML/Flux.jl/pull/600) for updated APIs for use in the updated GPU stack @maleadt
7. The Iris dataset can now be accessed as [Data.Iris](https://github.com/FluxML/Flux.jl/pull/652) thanks to @JoshuaWhittemore

Ongoing:

1. The major ongoing project is to [incorporate Zygote.jl](https://github.com/FluxML/Flux.jl/pull/669) as the AD. The work can also be tracked on Flux’s [zygote branch](https://github.com/FluxML/Flux.jl/tree/zygote).

[Tracker](https://github.com/FluxML/Tracker.jl) has now been moved into its own package as a step towards making Flux compatible with a larger number of AD libraries.

Please try out this release and file issues and regressions as necessary.

Please reach out if you wish to participate in the Google Summer of Code. The proposed projects list can be found [here](https://fluxml.ai/gsoc.html). But do feel free to come up with your own!

A huge thanks to everyone’s contributions with issues and PRs.

- Dhairya

---

<div class="post-metadata">

**Author:** ![Tamas\_Papp](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/tamas_papp/32/25949_2.png) [@Tamas\_Papp](https://discourse.julialang.org/u/Tamas_Papp)\
**Post date:** [March 25, 2019, 3:31pm UTC](https://discourse.julialang.org/t/ann-flux-v0-8/22305/2 "2019-03-25T15:31:30Z")

</div>

I don’t do ML, but use AD all the time (for optimization and Bayesian inference), so I am very happy that these excellent AD packages are decoupled from Flux, even if they share a lot of developers. This should make them easier to use for others, and may also benefit Flux users indirectly.