-
-
Notifications
You must be signed in to change notification settings - Fork 525
Closed
Description
It looks like pywt has a memory leak. When I process more than 200K small images (<10 KB each) a process takes all 16GB of my memory and stops (or slows down).
- Code
whash() function from imagehash lib:
JohannesBuchner/imagehash@da9386d
Reproduction code snippet:
res = {}
i = 0
for filename in onlyfiles:
fname = os.path.join(filedir, filename)
try:
image = PIL.Image.open(fname)
whash = imagehash.whash(image)
value = whash.hash.flatten()
res[int(filename[:-4])] = sum(1<<i for i, b in enumerate(value) if b)
except Exception as e:
print("Error. File " + fname + ": ", e )
Btw.. imagehash.phash() function doesn't use pywt and doesn't have this issue.
- Data
Image subset from Avito competition:
https://www.kaggle.com/c/avito-duplicate-ads-detection
Reactions are currently unavailable