search

Xtra-Computing / thundersvm

ThunderSVM: A Fast SVM Library on GPUs and CPUs
Apache License 2.0
1.56k stars 216 forks source link

Not even working for me in jupyter #140

Closed infoankit10 closed 5 years ago

infoankit10 commented 5 years ago

I am trying to use it on (5000, 3009) shape dataset in jupyter notebook. Jupyter kernel gets failed, although its working fine with sklearn svm.SVC method. Any quick solution? 

#### This code not working ####
start = time. time()
clf = SVC(kernel='linear')
clf.fit(X_train, Y_train)
end = time. time()
print(end-start)
#### This code working fine ####
start = time. time()
clf = svm.SVC(kernel='linear')
clf.fit(X_train, Y_train)
end = time. time()
print(end-start)

Actually, my overall data set is very large. Its shape is around (150000, 3009). So I am looking if your package can provide better performance.

Regards, Ankit Aggarwal

zeyiwen commented 5 years ago

We have tried thundersvm on jupyter nodebook, and it works fine... Please provide a minimum example for us to reproduce your issue.