Closed uuhh25 closed 2 years ago
all the training detail was put on the paper
all the training detail was put on the paper
thank u so much!
My pleasure But I have to say it seem not possible to re-performance the effective metric on datasets casia if the model was only trained on USCISI
My pleasure But I have to say it seem not possible to re-performance the effective metric on datasets casia if the model was only trained on USCISI
Yep, I met the situation you say and I try to figure out it. May you give some suggestions to tackle it?
If using the two branches’ output as the fusion‘s input when we train the fusion module? Or use the two branches’ label to train the fusion module. If someone can give me some suggestion, it is so amazing!