Service container exited with error "There appear to be 1 leaked semaphore objects to clean up at shutdown"

[SLIBM]

The service container (using image: primeqa/services:0.9.9-gpu) exited with error in the log like below:
(The tail of the log)

{"time":"2023-03-14 19:38:58,464", "name": "ReaderService", "level": "INFO", "message": "Applying 'ExtractiveReader' reader for query = 'what is the risk of taking these steps?' returns predictions = [[{'example_id': '0', 'span_answer_text': 'an interruption of the service is acceptable', 'span_answer': {'start_position': 1332, 'end_position': 1376}, 'confidence_score': 0.3760988641962656}, {'example_id': '0', 'span_answer_text': 'an interruption of the service', 'span_answer': {'start_position': 1332, 'end_position': 1362}, 'confidence_score': 0.36555731290584886}, {'example_id': '0', 'span_answer_text': 'interruption of the service is acceptable', 'span_answer': {'start_position': 1335, 'end_position': 1376}, 'confidence_score': 0.25834382289788543}]]"}
100%|██████████| 1/1 [00:00<00:00, 266.97it/s]/usr/lib/python3.8/multiprocessing/resource_tracker.py:216: UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown
  warnings.warn('resource_tracker: There appear to be %d '

I don't know the specific steps to reproduce this as it often exited without my notice, but I seem to observe this behavior more frequently if we switch the Reader model back and forth for experimentations.