Open drahnr opened 6 years ago
Could you post the latest snippet from git log --oneline
?
023ed5c (HEAD -> wip-dropout) fix/test: implement positive an negative tests for dropout
f212dd3 fix/warning: remove a mut from native helper
00f3464 feat/dropout: native implementation
6d0f02d (origin/wip-dropout) feat/test: add dropout test
f434280 fix/test: conv is not normalized by default
25a9eff fix/test: convolution is normalized
b6d360c feat/test: verify the results of convolution unit tests
9730c89 fix/pooling: swap padding and stride arguments in test
8534e71 chore/doc: update author and project name in CONTRIBUTING.md
7a45a3e chore/ci: remove travis related config
217be7a fix/dep: move from path to git dependencies
dd2faaa fix/dropout: add missing backend implementation for cuda
78d6059 feat/layers: add dropout layer
5a290b8 chore/rename: adjust urls to the new org
f9bb737 chore/rename: move to rcudnn crate
3702bd6 chore/rename: rename to coaster-nn
5a5e1a1 test/doc: make commented example compile
d6beed3 (origin/wip-native-max-pool-grad) feat/native: start hacking on backprop for maxpool
de50427 test/cross: add max pooling grad
4d6dc7c chore/warnings: remove mutable and unused backend instance
6da16b3 fix/cudnn: add unimplemented stances for future cudnn conv enums
d15ac01 test/cross: add cross test for softmax
4e8a5e3 docs/comments: supported operations table
d1fe941 chore/warnings: remove unused macros from native
3480b8a Merge pull request #10 from ratpoison-io/wip-decoupling
ecdd914 (origin/wip-decoupling) docs/example: use new API for examples in comments documentation
85c4bb9 feat/test: expand pooling test
c7d5cc0 fix/pooling: fix native pooling for negative numbers
952520c fix/test: add stride and one increment for result dimensions
b9758e7 refactor/tensor: backend decoupling and typed memory interface
2b968d3 docs/readme: update maintainer
9269374 chore/doc: update supported cudnn version stated
8dbb151 Merge pull request #6 from ratpoison-io/wip-cudnn5105
1124bb8 (origin/wip-cudnn5105) fix/test: correctly set result tensor dimensions for convolutional tests
a54362c feat/test: add pooling cross test
4ae37c0 feat/pooling: stride before padding argument
a16437c fix/warnings: mod/use related warnings
8d57009 chore/format: format feature table
9990189 test/conv: add first cross backend test for convolution
f49e9a4 fix/test: fix pooling test max grad check, fix dimensions of a few more
da0ba4d refactor/test: change mod helpers to deal with different backends
2329539 fix/winograd: make convolution algorithims for cuda exhaustive
902f696 chore/cudnn: update api to use latest v5.1 cudnn
c4f4ff0 Merge pull request #5 from ratpoison-io/wip-generic-cuda
0b5f107 (origin/wip-generic-cuda) fix/cuda: adjust to log softmax api change
74fbd5f fix/native: explicit type to fix compilation
0fcd84b fix/native: explicit type to fix compilation
b113d3f chore/format: run rustfmt for cuda mod
ad96a53 refactor/adapt: use new read write macros for tensor access
024b9b5 refactor/fallout: merge generic cuda patch, this is just to fix fallout
5cbc3a7 refactor/cuda: implement the plugin traits generically against the Cuda backend
Ah ok, I think it's because clog doesn't support section/component: description
it supports section(component): description
I misread your original message and thought foo/bar
was a custom section you'd added to your .clog.toml
Adding support for this wouldn't be hard amd would just be changing the regex, but I'd be worried about breaking anyone using foo/bar(component): description
custom headers.
That is alright, I was just confused, because I took over the project and it always used this syntax and it seemed to work until late 2015.
One could make the regex in a way that would either allow what/component
in names if or what(component)
. I'll draft something up.
^(?:([^:\/]+?)(?:\/([^:\(]*?)?)?|([^:\(]+?)(\(([^\)]*?)\))?):(.*)
this would be the new regex, https://regex101.com/r/LluL4e/1/tests