Building OpenVINO on Ubuntu

[Rainmanne]

I am trying to build it using the readme instructions and get stuck with the following:

Sub-Component builds

We’re now going to build whisper.cpp.

# OpenVINO
source /path/to/l_openvino_toolkit_ubuntu22_*_x86_64/setupvars.sh

# Libtorch
export LIBTORCH_ROOTDIR=/path/to/libtorch

Not sure what I am supposed to do here. I’ve got experience building stuff on Ubuntu in the past. Have not done it for a while. But still, it’s not clear to me what to do it.

[Rainmanne]

OK, I have figured it out. The only issue now is that the transcription plugin does not pick up the models. They are in /usr/local/lib/openvino-models but when I do Analyze/OpenVINO Whisper Transcription… in Audacity, the Whisper Model field is greyed out.