tencent-ml-images
tencent-ml-images copied to clipboard
over-fitting
when you train multi-labels classification on you dataset, whether there is serious overfitting? whether you novel loss function can resolve this problem?