Recently,i am obsessed in the few-shot learning,and try to extend it into other computer vision tasks.I'am surprised that the terms of support set and query set when training this network ,moreover,the mete-training method just like the "optimization to adapt,em MAML" but not at all.So,could you give some more details about the few-shot learning strategy?
Recently,i am obsessed in the few-shot learning,and try to extend it into other computer vision tasks.I'am surprised that the terms of support set and query set when training this network ,moreover,the mete-training method just like the "optimization to adapt,em MAML" but not at all.So,could you give some more details about the few-shot learning strategy?