Provides offline speech recognition features for the Freeswitch based on whisper_cpp.
People who are interested in this module, look at whisperd it'll may be more suitable.
Was capable to work with the first versions of whisper_cpp (not longer supported)
Capable to work with the latest version whisper_cpp (revelant on the date: 12.06.2024)
Before compile the module, you should build the whisper_cpp as a shared library and (possible) correct it's paths in the Makefile.
The issues: unfortunately i don't have enough time (now) to deeply dig what is going on, mb it's my mistake somewhere there.
But have the follogin (although the same time the whisperd works well):
2024-06-12 07:34:38.686528 99.73% [NOTICE] utils.c:115 transcribe samples=31360
terminate called after throwing an instance of 'std::length_error'
what(): cannot create std::vector larger than max_size()
Aborted (core dumped) ./bin/freeswitch -nf -nonat -nonatmap -nort<extension name="asr-tets">
<condition field="destination_number" expression="^(3222)$">
<action application="answer"/>
<action application="sleep" data="1000"/>
<action application="play_and_detect_speech" data="/tmp/test2.wav detect:whisper {lang=en}"/>
<action application="log" data="CRIT SPEECH_RESULT=${detect_speech_result}"/>
<action application="sleep" data="1000"/>
<action application="hangup"/>
</condition>
</extension>
# using freeswitch packages
apt update ; apt install libfreeswitch-dev libfreeswitch1 libfreeswitch1-dbg
cd mod_whisper_asr/sources
git clone https://github.com/ggerganov/whisper.cpp.git whisper
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ..
make -j
make install