Conversation
*******************************************************************************/ | ||
|
||
#include "ngraph/runtime/plaidml/plaidml_builder.hpp" | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
remove this line. Also <>
headers go before ""
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Done.
(FWIW: This pattern comes from the Google coding standards; the idea is that each header should be the first header included by some .cpp file, to ensure that each header is properly self-contained.)
vertexai::plaidml::application finalize() const; | ||
|
||
// Adds an input to the function. | ||
Function& add(Input input) &; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What is the problem being solved with returning &
and &&
? I haven't used this before.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The idea was to make it easy to use this
as either an lvalue and rvalue-reference, so that you could do either MyObject foo; foo.bar().baz(); return foo.finalize()
or return MyObject{}.bar().baz().finalize()
, depending on what reads best at the call site.
Should probably designate |
Other external dependencies are automatically downloaded and built though cmake configurations. |
We haven't released an SDK for PlaidML yet; it's currently just a Bazel rule that builds a tar file with the right contents. I think we'd like to have a non-Python binary release of some sort, but it's not our highest priority, and I think we'd need to think a bit more about what it'd look like. (E.g. on MacOS, it seems like we might want it to be a Framework.) It might be possible to download PlaidML and compile it via shelling out to Bazel. I suspect it'd take me a while to get that working, but it might be simple for someone with more expertise with cmake. It'll probably be easier after our next open-source release (hopefully soon); that should include a stable source tarball. (I very briefly thought about trying to talk nGraph devs into adding a Bazel build system, which would make it trivial to compile PlaidML as part of the build. That seems like a much longer conversation, though, and doesn't really help unless it becomes the only build system.) |
Used m_ prefix for members; removed trailing underscores Updated license headers Moved associated header inclusions to project blocks Wrapped comments to 100 chars Added missing newlines between functions Removed nested namespaces in operation implementations
FWIW -- I don't have access to the Jenkins server; anyone want to let me know what failed? :-/ |
@earhart I put in the correct (I think :/) access requests for you. Can take a look at the CI for you later this afternoon if the approval hasn't come through by then. |
Awesome; thanks! I'll keep checking it periodically. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks like @rkimballn1 already got it. The problem was the source code formatting check. You can run |
No description provided.