开放
Weights shape not validated against kernel, channels
good first issuehelp wanted
5 条评论1 个反应0 负责人
议题
开放
ConvTranspose with padding on cpu throws exception
bughelp wanted
1 条评论0 个反应0 负责人
开放
Add docs note about saving/loading models with anonymous functions
documentationhelp wanted
9 条评论1 个反应0 负责人
开放
Separate nightly CI from posted CI badge
discussiongood first issuetesting
2 条评论0 个反应0 负责人
开放
reducing invalidations
help wanted
1 条评论0 个反应0 负责人
开放
Functional equivalents missing for some layers
enhancementgood first issue
3 条评论0 个反应0 负责人
开放
More robust testing for AlphaDropout
good first issuehelp wantedtesting
6 条评论0 个反应0 负责人
开放
Overloading Flux.batch for custom eltypes
enhancementgood first issuehelp wanted
13 条评论0 个反应0 负责人
开放
Figure out consistent code style
discussionhelp wanted
9 条评论0 个反应0 负责人
开放
Benchmark optimizer latency on GPU
cudahelp wantedperformance
0 条评论0 个反应0 负责人
开放
Allow BatchNorm training on CUDA with `track_stats=false`
good first issuehelp wanted
4 条评论0 个反应0 负责人
开放
PyTorch feature parity
help wanted
94 条评论33 个反应0 负责人
开放
Shape-propagating Chain
discussionenhancementhelp wanted
8 条评论8 个反应0 负责人
开放
'cell_clip' and 'proj_clip' for Flux: LSTM
enhancementhelp wanted
1 条评论0 个反应0 负责人
开放
Optional batch dimension for convolutions
help wanted
3 条评论0 个反应0 负责人
开放
Provide aliases for unicode fields and keywords
good first issuehelp wanted
5 条评论0 个反应0 负责人
开放
Does Flux support N-dimensional convolutions, where N>3?
enhancementhelp wanted
20 条评论0 个反应0 负责人
开放
Add max/mean unpooling operations
enhancementgood first issuehelp wanted
5 条评论6 个反应0 负责人
开放
ConvRNN Cell and ConvLSTM Cell
RNNhelp wanted
5 条评论0 个反应0 负责人