training and test problem

[songjianwu]

Thank you for sharing your project.
when I apply PointSIFT into our data,it is normal in training phase, however, in test phase,the loss is big and the accuracy is low ,could you give me some suggestion? thank you.

[braveSunflower]

Thank you for sharing your project.
when I apply PointSIFT into our data,it is normal in training phase, however, in test phase,the loss is big and the accuracy is low ,could you give me some suggestion? thank you.

Hello ,could you please share the test.py with me ,my email is [email protected],thanks a lot!

[AnkyZhang]

test.py

Hi， Have you got the test.py?

[braveSunflower]

I'm sorry to answer your letter until this time , unfortunately,I haven't got the test.py file ,but some friends told me that the results given in paper came from the train.py directly,just as the Pointnet++. 在 2020-01-15 16:12:27，"AnkyZhang" <[email protected]> 写道： test.py Hi， Have you got the test.py? — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[China-First-KY]

Thank you for sharing your project.
when I apply PointSIFT into our data,it is normal in training phase, however, in test phase,the loss is big and the accuracy is low ,could you give me some suggestion? thank you.

Hello, I see your comments. I just made pointsift. Can you give me a contact information, wechat, QQ or email? Thank you again. I really need help!

[China-First-KY]

Thank you for sharing your project.
when I apply PointSIFT into our data,it is normal in training phase, however, in test phase,the loss is big and the accuracy is low ,could you give me some suggestion? thank you.

Hello, I see your comments. I just made pointsift. Can you give me a contact information, wechat, QQ or email? Thank you again. I really need help!