I'm running some tests on Orange Pi 5 and noticed that Vosk is running on one of the efficiency cores. OPi5 has 8 cores (4 performance, 4 efficiency).
Its still faster than realtime 😄, so its ok for streaming audio use-cases, but for batch transcriptions it could be much faster.
Can we tweak the code somehow to select a performance core instead?
I'm running some tests on Orange Pi 5 and noticed that Vosk is running on one of the efficiency cores. OPi5 has 8 cores (4 performance, 4 efficiency). Its still faster than realtime 😄, so its ok for streaming audio use-cases, but for batch transcriptions it could be much faster. Can we tweak the code somehow to select a performance core instead?