Heavy data,
please load training and test sets from


train : 6.9 G:
X = http://sabiod.univ-tln.fr/workspace/challenges/DOCC10/DOCC10_samples_train.npy
Y = http://sabiod.univ-tln.fr/workspace/challenges/DOCC10/DOCC10_train.csv

test : 1.3 G:
X = http://sabiod.univ-tln.fr/workspace/challenges/DOCC10/DOCC10_samples_test.npy
Y = http://sabiod.univ-tln.fr/workspace/challenges/DOCC10/DOCC10_test.csv


The first 2000 rows of the test set is the public data set for the benchmark. The other rows of the test set is the private test set.