Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / FluxML/NNlibCUDA.jl issues and pull requests

#72 - Increase CUDA.jl compat to 5.0.0?

Issue - State: open - Opened by johnbcoughlin 8 months ago - 2 comments

#71 - NNlib 0.9

Issue - State: closed - Opened by bjarthur about 1 year ago - 3 comments

#70 - Adding EnzymeRules

Issue - State: open - Opened by sethaxen about 1 year ago - 7 comments
Labels: enhancement

#69 - type-piracy of materialize

Issue - State: open - Opened by vtjnash about 1 year ago - 1 comment
Labels: good first issue

#68 - Float16 CUDA `conv` broken on 5D tensors

Issue - State: open - Opened by nikopj over 1 year ago - 8 comments
Labels: bug

#66 - rework batchnorm

Pull Request - State: open - Opened by CarloLucibello over 1 year ago

#65 - CompatHelper: bump compat for "CUDA" to "4.0"

Pull Request - State: closed - Opened by github-actions[bot] over 1 year ago

#64 - SVectors on GPU CuArray cannot use index from CuArray

Issue - State: open - Opened by AhmedSalih3d over 1 year ago - 2 comments

#63 - Faster `softmax`

Pull Request - State: closed - Opened by mcabbott over 1 year ago

#62 - overload `NNlib._rng_compat_array`

Pull Request - State: closed - Opened by mcabbott over 1 year ago

#61 - Maxpool misbehaving in some edge cases

Issue - State: open - Opened by LlewellynS96 over 1 year ago - 4 comments
Labels: bug

#60 - Add CUDA tests for `dropout`

Pull Request - State: closed - Opened by mcabbott over 1 year ago - 1 comment

#59 - added fold/unfold and gpu tests

Pull Request - State: closed - Opened by nikopj over 1 year ago - 8 comments

#58 - Updates for CUDA.jl 4.

Pull Request - State: closed - Opened by maleadt over 1 year ago - 1 comment

#57 - `relu` propagates NaN on CPU but not on GPU

Issue - State: open - Opened by ericphanson almost 2 years ago - 7 comments
Labels: bug, good first issue, help wanted

#56 - Fix versions in Buildkite CI

Pull Request - State: closed - Opened by ToucheSir almost 2 years ago - 1 comment

#55 - Move `ctc_loss` from Flux to NNlibCUDA

Pull Request - State: closed - Opened by mcabbott almost 2 years ago - 2 comments

#54 - print/convert batchedadjtrans over cuarray

Pull Request - State: closed - Opened by chengchingwen about 2 years ago

#53 - save allocs during algorithm search

Pull Request - State: closed - Opened by maxfreu about 2 years ago - 5 comments

#52 - Gather throw error on cpu and gpu array inputs

Pull Request - State: open - Opened by yuehhua about 2 years ago - 4 comments

#51 - Add checkbounds for gather and support empty source array

Pull Request - State: open - Opened by yuehhua about 2 years ago - 33 comments

#50 - upper bound for NNlib

Pull Request - State: closed - Opened by CarloLucibello about 2 years ago - 7 comments

#49 - get rid of duplicate code in upsampling code by using dispatch

Pull Request - State: closed - Opened by maxfreu about 2 years ago - 5 comments

#48 - Restrict element type of activation overrides to CUDNN datatypes

Pull Request - State: closed - Opened by DomCRose about 2 years ago

#46 - conv on complex numbers

Issue - State: open - Opened by ToucheSir about 2 years ago
Labels: enhancement, help wanted

#45 - Missing Identity Activation Import

Pull Request - State: closed - Opened by avik-pal over 2 years ago - 7 comments

#44 - update `softmax` tests to match NNlib 0.8.3

Pull Request - State: closed - Opened by mcabbott over 2 years ago - 4 comments

#43 - specialize on op in scatter

Pull Request - State: closed - Opened by CarloLucibello over 2 years ago

#42 - No frule for some activations

Issue - State: closed - Opened by RS-Coop over 2 years ago - 7 comments

#41 - Do we need src/cudnn/activations.jl?

Issue - State: open - Opened by ToucheSir over 2 years ago

#40 - Use new `PoolDims` constructor for NNlib v0.8

Pull Request - State: closed - Opened by DhairyaLGandhi over 2 years ago

#39 - CompatHelper: bump compat for "NNlib" to "0.8"

Pull Request - State: closed - Opened by github-actions[bot] over 2 years ago

#38 - Update NNlib compat to 0.8

Pull Request - State: closed - Opened by darsnack over 2 years ago - 1 comment

#37 - NNlibCUDA Heisenbug in conv! with nonzero beta

Issue - State: closed - Opened by jw3126 about 3 years ago - 7 comments

#36 - track_stats option for batchnorm

Pull Request - State: closed - Opened by CarloLucibello over 2 years ago - 6 comments

#35 - Interesting resources for optimizing various ops

Issue - State: open - Opened by ToucheSir over 2 years ago

#34 - Add support for asymmetric padding for convlayers

Pull Request - State: closed - Opened by DrChainsaw over 2 years ago - 6 comments

#33 - Support gpu scatter/gather with CartesianIndex

Pull Request - State: closed - Opened by yuehhua over 2 years ago - 1 comment

#32 - add support for half precision gemm

Pull Request - State: open - Opened by bjarthur over 2 years ago - 6 comments

#31 - Add CUDA kernels for grid sampling

Pull Request - State: closed - Opened by pxl-th over 2 years ago - 2 comments

#30 - Slow ∇softmax! compared with generic version.

Issue - State: open - Opened by jumerckx over 2 years ago - 15 comments
Labels: enhancement, help wanted

#29 - gpu `scatter` with `CartesianIndex` not supported

Issue - State: closed - Opened by chengchingwen almost 3 years ago - 2 comments

#28 - remove unused kernels

Pull Request - State: closed - Opened by CarloLucibello almost 3 years ago

#27 - introduce linear upsampling

Pull Request - State: closed - Opened by maxfreu almost 3 years ago - 2 comments

#26 - Release Patch

Pull Request - State: open - Opened by DhairyaLGandhi almost 3 years ago

#25 - Test fixes

Pull Request - State: closed - Opened by maleadt almost 3 years ago - 1 comment

#24 - BatchNorm causes error during gradient computation

Issue - State: open - Opened by pxl-th almost 3 years ago - 7 comments

#23 - introduce trilinear upsampling

Pull Request - State: closed - Opened by maxfreu almost 3 years ago - 1 comment

#22 - add depthwise_conv* overloads for CUDA

Pull Request - State: open - Opened by DhairyaLGandhi almost 3 years ago

#21 - remove some old cruft

Pull Request - State: closed - Opened by CarloLucibello almost 3 years ago

#20 - Bump patch and check for failures with Julia 1.7-nightly

Pull Request - State: closed - Opened by DhairyaLGandhi about 3 years ago - 1 comment

#19 - Use `CUDA.@atomic`

Pull Request - State: closed - Opened by darsnack about 3 years ago - 1 comment

#18 - Add links to readme

Pull Request - State: closed - Opened by mcabbott about 3 years ago

#17 - Test `softmax` with varying dimensions

Pull Request - State: closed - Opened by ToucheSir about 3 years ago

#16 - Replace workspace macro with do-block

Pull Request - State: closed - Opened by akashkgarg about 3 years ago - 4 comments

#15 - Update scatter.jl, specify which `@atomic`

Pull Request - State: closed - Opened by domschl about 3 years ago - 2 comments

#14 - Specify which @atomic required

Pull Request - State: closed - Opened by colbec about 3 years ago - 2 comments

#13 - Support scatter for CUDA gradient

Pull Request - State: closed - Opened by yuehhua about 3 years ago - 4 comments

#12 - Support gather for cuda gradient

Pull Request - State: closed - Opened by yuehhua about 3 years ago

#11 - `batched_gemm!` with CuArrays doesn't work

Issue - State: closed - Opened by bjarthur about 3 years ago - 3 comments

#10 - CUDNN bad param for backward pass with Float16

Issue - State: open - Opened by DrChainsaw about 3 years ago - 4 comments

#9 - Add group support for convolutions

Pull Request - State: closed - Opened by pxl-th about 3 years ago - 7 comments

#8 - Gather for CUDA support

Pull Request - State: closed - Opened by yuehhua about 3 years ago - 6 comments

#7 - Fix use of convdims.

Pull Request - State: closed - Opened by maleadt about 3 years ago - 3 comments

#6 - Add default value to convdims

Pull Request - State: closed - Opened by DrChainsaw about 3 years ago - 3 comments

#5 - introduce trilinear upsampling

Pull Request - State: closed - Opened by maxfreu about 3 years ago - 6 comments

#4 - TagBot trigger issue

Issue - State: closed - Opened by JuliaTagBot about 3 years ago - 20 comments

#3 - Add a License

Pull Request - State: closed - Opened by DhairyaLGandhi about 3 years ago

#2 - Add TagBot and CompatHelper

Pull Request - State: closed - Opened by DhairyaLGandhi about 3 years ago

#1 - Add scatter for CUDA support

Pull Request - State: closed - Opened by yuehhua about 3 years ago - 15 comments