Pull Requests FluxML/Flux.jl

CompatHelper: bump compat for CUDA in [weakdeps] to 6, (keep existing compat)

#2672 opened 2026-04-11 00:34 by github-actions[bot]

CompatHelper: bump compat for cuDNN in [weakdeps] to 6, (keep existing compat)

#2671 opened 2026-04-11 00:34 by github-actions[bot]

fix stack overflow when passing empty struct

#2670 opened 2026-04-02 11:31 by hardik-xi11

Bump codecov/codecov-action from 5 to 6 dependencies github_actions

#2668 opened 2026-03-30 22:54 by dependabot[bot]

Unmark Bilinear and MultiHeadAttention Enzyme tests as broken on Julia 1.12

#2666 opened 2026-03-26 08:30 by Copilot

test mooncake on CUDA gpu

#2660 opened 2026-03-16 16:01 by CarloLucibello

Enable FLUX_TEST_REACTANT in AMD pipeline configuration

#2646 opened 2026-01-02 18:07 by wsmoses

add opencl support

#2635 opened 2025-12-08 08:40 by VarLad

Adding `summarise_array` to show.jl

#2593 opened 2025-03-01 13:06 by murrellb

Add a "gradient zoo" page? documentation gradients enzyme

#2447 opened 2024-05-14 02:43 by mcabbott

Add repeat vector

#2409 opened 2024-03-24 18:52 by Jafagervik

freezingdocs documentation

#2397 opened 2024-03-13 05:07 by isentropic

Added test for Alpha Dropout

#2367 opened 2023-12-30 18:56 by Bhavay-2001

Correct counting of shared parameters in `Base.show`.

#2335 opened 2023-09-14 09:15 by codetalker7

Make BatchNorm twice-differentiable

#2255 opened 2023-05-08 18:47 by alexrosen45

Broken link fixes

#2209 opened 2023-03-13 23:36 by Vaibhav-sa30

Propose accuracy functions

#2181 opened 2023-02-05 16:06 by skyleaworlder

Add LPNormPool

#2166 opened 2023-01-24 13:55 by skyleaworlder

accept integer labels in (logit)crossentropy

#2141 opened 2022-12-26 14:56 by CarloLucibello

[Feature] Logit focal loss

#2138 opened 2022-12-20 19:11 by josephsdavid

Add a macro `@throttle` enhancement

#2120 opened 2022-11-21 16:01 by mcabbott

Improve README example

#2108 opened 2022-11-10 13:54 by MilesCranmer

Remove type restrictions on normalization layers' scale and bias breaking run downstream test

#2099 opened 2022-11-03 18:42 by darsnack

Make `loss(f,x,y) == loss(f(x), y)`

#2090 opened 2022-10-18 16:35 by mcabbott

Make `outputsize` work with `Embedding`

#2088 opened 2022-10-16 14:59 by mcabbott

Speed up `show` for gpu models

#2066 opened 2022-09-26 17:45 by mcabbott

Improve type stability of LayerNorm and Dropout performance

#2005 opened 2022-06-23 05:17 by ToucheSir

Attempt to write a type stable `rrule(Chain{Vector})` help wanted gradients

#2003 opened 2022-06-20 06:17 by ToucheSir

cleanup Flux.Losses documentation documentation

#1930 opened 2022-04-06 00:00 by CarloLucibello

Use LogExpFunctions for losses

#1866 opened 2022-02-05 21:29 by ToucheSir