Replace `multiprocessing.dummy.Pool()` with `concurrent.futures.ThreadPoolExecutor()` so whisper_s2t instance can run separately with `multiprocessing.Process()`

shashikg / WhisperS2T

An Optimized Speech-to-Text Pipeline for the Whisper Model Supporting Multiple Inference Engine

MIT License

315 stars 32 forks source link

Replace `multiprocessing.dummy.Pool()` with `concurrent.futures.ThreadPoolExecutor()` so whisper_s2t instance can run separately with `multiprocessing.Process()` #74

Open kirillsaidov opened 1 month ago

kirillsaidov commented 1 month ago

I tried loading whisper model and running transcription in a separate multiprocessing.Process(), but failed. The process just froze unresponsive when loading audio to memory with multiprocessing.dummy.Pool(). Replaced with a safer concurrent.futures.ThreadPoolExecutor() and it works now!