Hi,
running the transcription with the app in foreground, when passing to background I immediately got an exception in whisper.cpp.
I am using whisper.cpp with METAL, so I'm running on GPU.
I suppose that this is the problem: https://developer.apple.com/documentation/metal/gpu_devices_and_work_submission/preparing_your_metal_app_to_run_in_the_background
I'm using the latest commit available, 940de9d, and the problem occurs here, in file ggml-metal:
What are the possible strategies to not going in error or to pause the inference and get back to the transcription progress when the app came back to foreground ?
Thank you so much for your work!