FluxML/Flux.jl

Issues

• Flux.cpu and Flux.gpu no longer move data on views

  #2553 opened 7 days ago by piever
  3

• Enzyme gradient example broken

  #2554 opened 8 days ago by lassepe
  5

• Stacked RNN in Flux.jl?

  #2452 opened 9 days ago by NeroBlackstone
  1

• cell output is not clearly distinguishable from the state

  #2548 opened 12 days ago by CarloLucibello
  3

• feature requests for RNNs

  #2514 opened 2 months ago by CarloLucibello
  9

• `Flux.@functor` is broken on 0.15

  #2545 opened 13 days ago by mcabbott
  1

• display fieldnames in show of a container

  #2543 opened 16 days ago by CarloLucibello
  2

• cuda gpu memory usage increasing in time

  #2523 opened a month ago by CarloLucibello
  3

• device movement behavior inconsistent

  #2513 opened 18 days ago by chengchingwen
  2

• broken image in the quickstart

  #2530 opened 18 days ago by CarloLucibello
  1

• mark as public any non-exported but documented interface

  #2518 opened 20 days ago by CarloLucibello
  1

• `Flux.params` is broken

  #2533 opened 20 days ago by mcabbott
  0

• Consider making the `:expand` option the default in `@layer`

  #2531 opened 20 days ago by CarloLucibello
  7

• missing codecov token in buildkite

  #2517 opened 2 months ago by CarloLucibello
  1

• `gpu` should warn if cuDNN is not installed

  #2440 opened a month ago by CarloLucibello
  3

• use DocumenterVitepress.jl for documentation

  #2519 opened 2 months ago by CarloLucibello
  1

• Implementation of `AdamW` differs from PyTorch

  #2433 opened 2 months ago by dpaetzel
  11

• Add option to throw error on passing wrong precision floats to layers

  #2454 opened 7 months ago by lassepe
  5

• is `Flux.huber_loss` type-unstable ?

  #2459 opened 6 months ago by filchristou
  1

• DifferentiationInterface testing

  #2469 opened 5 months ago by gdalle
  6

• Depthwise convolutions produce a large number of allocations

  #2508 opened 2 months ago by JoshuaBillson
  1

• test Enzyme gpu support

  #2445 opened 7 months ago by CarloLucibello
  0

• Collecting PyTorch -> Flux migration notes

  #2410 opened 9 months ago by BioTurboNick
  2

• deprecate Flux.params

  #2413 opened 2 months ago by CarloLucibello
  7

• Adding Simple Recurrent Unit as a recurrent layer

  #2408 opened 2 months ago by Jafagervik
  1

• Flux has no Lecun Normalization weight init function?

  #2491 opened 2 months ago by neoareslinux
  2

• ERROR: UndefVarError: `ADAM` not defined in `Main` in flux

  #2507 opened 2 months ago by Liuxiu1512
  1

• Weights shape not validated against kernel, channels

  #2506 opened 2 months ago by BioTurboNick
  4

• any way to do causal pad in Flux ?

  #2505 opened 2 months ago by neoareslinux
  4

• Problem with documentation

  #2485 opened 2 months ago by vsoler
  0

• Potential bug of RNN training flow

  #2455 opened 2 months ago by Bonjour-Lemonde
  1

• Cannot take `gradient` of L2 regularization loss

  #2441 opened 2 months ago by Vilin97
  1

• Zygote fails to differentiate through Flux.params on julia v0.11

  #2497 opened 2 months ago by CarloLucibello
  2

• The dependency error about `Flux->FluxMPIExt` occurs when updating to Julia 1.11

  #2490 opened 2 months ago by pinkie-ljz
  1

• use MLDataDevices.jl?

  #2482 opened 2 months ago by CarloLucibello
  0

• Performance issues when switching from local machine to HPC cluster.

  #2484 opened 3 months ago by SGHoekstra
  7

• Model saved under Flux v0.14.16 does not load on v0.14.17

  #2476 opened 4 months ago by cirobr
  6

• ConvTranspose errors with symmetric non-constant pad

  #2424 opened 5 months ago by piever
  0

• Requires deprecated cuNN.jl package

  #2470 opened 5 months ago by JakobAsslaender
  1

• ConvTranspose with padding on cpu throws exception

  #2465 opened 6 months ago by DrChainsaw
  1

• Create a flag to use Enzyme as the AD in training/etc.

  #2443 opened 6 months ago by wsmoses
  14

• Can't load a Fluxml trained & saved model. Getting ERROR: CUDA error: invalid device context (code 201, ERROR_INVALID_CONTEXT)

  #2461 opened 6 months ago by georgegrosu1
  1

• why is my `withgradient` type unstable ?

  #2456 opened 6 months ago by filchristou
  1

• Enable github Discussions

  #2451 opened 7 months ago by NeroBlackstone
  0

• Enzyme fails with MultiHeadAttention layer

  #2448 opened 7 months ago by mashu
  13

• test Enzyme gradient for loss functions

  #2444 opened 7 months ago by CarloLucibello
  0

• Dense layers with shared parameters

  #2432 opened 7 months ago by pagnani
  5

• SamePad() for even sized filters.

  #2431 opened 8 months ago by JohnAshburner
  0

• tests are failing due to ComponentArrays

  #2411 opened 9 months ago by CarloLucibello
  2

• Significant time spent moving medium-size arrays to GPU, type instability

  #2414 opened 9 months ago by BioTurboNick
  10