Task dependent Deep LDA pruning of neural networks

Qing Tian, Tal Arbel, James J. Clark

With deep learning's success, a limited number of popular deep nets have been widely adopted for various vision tasks. However, this usually results in unnecessarily high complexit…