it seems that the face_recognition does not release memory properly
- face_recognition version:1.2.3
- Python version:3.7.3
- Operating System:win10
Description
it seems that the face_recognition does not release memory properly, after complete the process, it still holds the GPU memory. Any advice?
What I Did
#start #... RAM usage: 38.1% 6231M/16349M; CPU usage: 35.4% #... GPU#0: total memory 2048 M, used 1073 M 52.4 %, free 974 M
faces_locations= face_recognition.face_locations(frame,number_of_times_to_upsample=1, model='cnn') face_encodings = face_recognition.face_encodings(frame, faces_locations) result = face_recognition.compare_faces([face_encodings_kown], face_encodings[0], tolerance=0.6)
# ... RAM usage: 43.3% 7072M/16349M; CPU usage: 12.3%
#... GPU#0: total memory 2048 M, used 1986 M 96.98 %, free 61 M
faces_locations= face_recognition.face_locations(frame,number_of_times_to_upsample=1, model='cnn') face_encodings = face_recognition.face_encodings(frame, faces_locations) result = face_recognition.compare_faces([face_encodings_kown], face_encodings[0], tolerance=0.6)
#... RAM usage: 43.6% 7130M/16349M; CPU usage: 11.2%
#... GPU#0: total memory 2048 M, used 1916 M 93.55 %, free 131 M
#end # ... RAM usage: 43.9% 7169M/16349M; CPU usage: 16.5% # ... GPU#0: total memory 2048 M, used 1916 M 93.55 %, free 131 M
Did you find any solution regarding this memory issue?
Im facing same problem , any solution please?