Add FL Decoder / KenLM integration to build process #2078
Conversation
mthrok
force-pushed
the
decoder-integration
branch
from
60d932f to
2243480
Compare
|
@mthrok has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
@mthrok has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
| ) | ||
| endif() | ||
|
|
||
| # TODO: Add libtorchaudio_decoder |
There was a problem hiding this comment.
what is necessary for adding libtorchaudio_decoder?
There was a problem hiding this comment.
The definition of TORCHAUDIO_LIBRARY that is expected to be used from C++ integration via CMake. To make the decoder available in C++, we need to add libtorchaudio_decoder to it. However, the current way is not entirely correct, so we need to revise it.
| if (BUILD_CTC_DECODER) | ||
| set( | ||
| LIBTORCHAUDIO_DECODER_SOURCES | ||
| decoder/src/decoder/Decoder.h |
There was a problem hiding this comment.
we should be able to remove the .h header files from this list
There was a problem hiding this comment.
Thanks for pointing. Let me follow-up on this in the subsequent PR so that I can land this without alerting the fbcode's change-after-approve system during the weekend.
Summary: After all the C++ code from pytorch#2072 are added, this commit will enable decoder/KenLM integration in the build process. Pull Request resolved: pytorch#2078 Reviewed By: carolineechen Differential Revision: D33198183 Pulled By: mthrok fbshipit-source-id: 9d7fa76151d06fbbac3785183c7c2ff9862d3128
3 participants
After all the C++ code from #2072 are added, this commit will enable decoder/KenLM integration in the build process.