Hi there,
I'm currently running a Django webapp that uses OpenCV (cv2) and Numpy for image processing/face detection. One of my Caffe files is 500mb and everytime I hit my API it has to load these files making the request several seconds to execute. I was wondering if there's a way to load these files up into memory / cache the files. When I try with the normal caching methods I get a pickling error.
My code to load up my Caffe files is:
def load_face_net():
THIS_FOLDER = os.path.dirname(os.path.abspath(__file__))
FACE_PROTO_PATH = os.path.join(THIS_FOLDER, FACE_PROTO)
FACE_MODEL_PATH = os.path.join(THIS_FOLDER, FACE_MODEL)
# face_net = cache.get('face_net')
# if face_net is None:
face_net = cv2.dnn.readNetFromCaffe(FACE_PROTO_PATH, FACE_MODEL_PATH)
# cache.set('face_net', face_net, 3600)
return face_net
def load_gender_net():
THIS_FOLDER = os.path.dirname(os.path.abspath(__file__))
GENDER_MODEL_PATH = os.path.join(THIS_FOLDER, GENDER_MODEL)
GENDER_PROTO_PATH = os.path.join(THIS_FOLDER, GENDER_PROTO)
# gender_net = cache.get('gender_net')
# if gender_net is None:
gender_net = cv2.dnn.readNetFromCaffe(GENDER_MODEL_PATH, GENDER_PROTO_PATH)
# cache.set('gender_net', gender_net, 3600)
return gender_net
varand
|
29
posts
|
fjl
|
4352
posts
|
PythonAnywhere staff
|
