kfzyqin / Research-Softmax-with-Mutual-Information

Full implementation of the paper "Rethinking Softmax with Cross-Entropy: Neural Network Classifier as Mutual Information Estimator".

kfzyqin/Research-Softmax-with-Mutual-Information Issues

Did you notice that the loss in the training process is negative ?
Closed 4 years ago3
Why do you divide the rtn[1] by 100 ?
Closed 5 years ago4