could you give some tips for building a fully-connected network with MAML++

[HeegerGao]

Hi, I want to use MAML++ to train a simple three-layer fully-connected network. But I found that all your config files are designed for the image input, in which I mean they contain parameters like "channels, kernel_size, padding" and so on. Besides, there is no existing model for FC networks in the few_shot_learning_system.py.

Could you write a config file and a complete process for running a fully-connected network with MAML++? Or could you please give some guidence on how to rewrite and use them? Thank you very much!