Hi I am using latest code for creating my own scorer, generate_lm.py script ran successfully and created vocab file and lm.binary too . My problem is since generate_pacakge.py has been rewritten to generate_scorer_package.cpp, how can make use of it to package and create the .scorer file?
I tried using g++ make it executable and pass the arguments.
But it is giving me error , No such file or directory #include "absl/types/optional.h"
Okay got you @othiele . Thank you !
But if at all I have to use the generate_scorer_package.cpp file to package the scorer , could you please point me to some documentation related to that?
lissyx
((slow to reply) [NOT PROVIDING SUPPORT])
5
Just follow the build instructions and use //native_client:generate_scorer_package instead of //native_client:libdeepspeech.so
lissyx
((slow to reply) [NOT PROVIDING SUPPORT])
7
When cloning the release codebase, what specifically do I git clone?
The documentation at https://deepspeech.readthedocs.io/en/v0.7.4/TRAINING.html just says to git clone https://github.com/mozilla/DeepSpeech. Won’t this pull the entire codebase?
What tags do we specify when cloning to get release 0.7.4 code only?