Awesome Open Source
Awesome Open Source


code release for "Towards Discriminability and Diversity: Batch Nuclear-norm Maximization under Label Insufficient Situations" ( CVPR2020 oral)

One-sentence description

We prove in the paper that Batch Nuclear-norm Maximization (BNM) could ensure the prediction discriminability and diversity, which is an effective method under label insufficient situations.


One line code under Pytorch and Tensorflow

Assume X is the prediction matrix. We could calculate BNM loss in both Pytorch and Tensorflow, as follows:


  1. Direct calculation (Since there remains direct approach for nuclear-norm)
L_BNM = -torch.norm(X,'nuc')
  1. Calculation by SVD
L_BNM = -torch.sum(torch.svd(X, compute_uv=False)[1])


L_BNM = -tf.reduce_sum(tf.svd(X, compute_uv=False))


We apply BNM to domain adaptation (DA) in DA, unsupervised open domain recognition (UODR) in UODR and semi-supervised learning (SSL) in SSL.

Training instructions for DA, UODR and SSL are in the in DA, UODR and SSL respectively.


If you use this code for your research, please consider citing:

author = {Cui, Shuhao and Wang, Shuhui and Zhuo, Junbao and Li, Liang and Huang, Qingming and Tian, Qi},
title = {Towards Discriminability and Diversity: Batch Nuclear-Norm Maximization Under Label Insufficient Situations},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2020}

Supplementary could be found in Google driver and baidu cloud (z7yt).



If you have any problem about our code, feel free to contact

or describe your problem in Issues.

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
python (50,867
domain-adaptation (58
transfer (28

Find Open Source By Browsing 7,000 Topics Across 59 Categories