jhjacobsen/invertible-resnet

training parameters for SVHN/other datasets

liangbright opened this issue · 1 comments

liangbright commented

Dear authors,

I am training the model on SVHN with coeff=0.3, after 200 epochs, many singular values are still larger than 1.
Are there any other parameters to be adjusted?

Best,
Liang

jhjacobsen commented

Try adjusting the number of power iterations for spectral normalization