Open Chenly1993 opened 7 years ago
Thank you for your contribution! Now ，I want to use it for 256 dimensional labels classification. Each label is 0 or 1. But the loss of training is strange. It will sharp drop at first. Then it will keep almost no changed.
Thank you for your contribution! Now ，I want to use it for 256 dimensional labels classification. Each label is 0 or 1. But the loss of training is strange. It will sharp drop at first. Then it will keep almost no changed.
