provide a replacement for pickle to work efficiently on arbitrary Python objects containing large data, in particular large numpy arrays

As of Python 3.8 and numpy 1.16, pickle protocol 5 introduced in PEP 574 supports efficient serialization and de-serialization for large data buffers

joblib.load() should therefore never be used to load objects from an untrusted source

import joblib

l = list(range(10))

f = open('temp.joblib', 'wb')

# serialization
joblib.dump(l, f)

f.close()
		

f = open('temp.joblib', 'rb')

# deserialization
print(joblib.load(f))

f.close()
        

l = list(range(10))

# serialization
joblib.dump(l, 'temp.pkl')

# deserialization
joblib.load('temp.pkl')
		

# support 'z', ‘gzip’, ‘bz2’, ‘lzma’ and ‘xz’:

l = list(range(10))

# serialization
joblib.dump(l, 'temp.z')

# deserialization
joblib.load('temp.z')
		

Document