A Softmax-free Loss Function Based on Predefined Optimal-distribution of Latent Features for Deep Learning Classifier

Open in new window