Docker gunicorn implementation stuck to transcribe in flask environment

[danirisdiandita]

when i run whisper model using flask app in gunicorn (in Docker). The transcription didn't work

from faster_whisper import WhisperModel 
class WhisperTranscriber(): 
    def __init__(self) -> None:
        self.faster_whisper = WhisperModel('/app/apps/model/whisper_medium', device='cpu', compute_type="int8")
    def transcribe(self, filepath): 
        # this part below stuck forever when run transcribe 
        segments, info = self.faster_whisper.transcribe(filepath, beam_size=5, word_timestamps=True)
        for seg in segments: 
             print(seg.text)

however, when i run the script in a docker exec -it <container_name> /bin/bash and run python script the inference works well

the gunicorn config gunicorn-cfg.py is shown below

workers = 1
accesslog = "-"
loglevel = "debug"
capture_output = True
enable_stdio_inheritance = True

and the command when running docker

gunicorn --timeout 7200 --certfile=server.crt --keyfile=server.key --config gunicorn-cfg.py -b 0.0.0.0:5005 run:app