After transcription, there is still occupation of GPU memory

[lesrose]

`from faster_whisper import WhisperModel

model = WhisperModel("/data/ai-model/faster-whisper-large-v3")
segments, info = model.transcribe("/data/resource-data/2024-06-19/gpu1/b5fa231dfc7a462d8c91dbbc36cd7c72.mp3")
for i in range(10):
time.sleep(10)`
Why is there still gpu memory usage during sleep

[toanhuynhnguyen]

You should implement this before the time.sleep(10):

del model
gc.collect()
for i in range(10):
    time.sleep(10)

gc is imported as below:

import gc