goldblum / AdversarialQuerying

A PyTorch implementation of the method found in "Adversarially Robust Few-Shot Learning: A Meta-Learning Approach"
MIT License
49 stars 10 forks source link

about maml-adml #2

Closed Herb2333 closed 4 years ago

Herb2333 commented 4 years ago

Thank you for your code and paper! However,why remove maml-adml files,i found comparative experiment on this two models in your paper,but there are not maml-adml files in it.

goldblum commented 4 years ago

The MAML code is very different than the rest of this repo, so it would be like having two repos in one, and MAML is very far from state-of-the-art right now. We wanted a clean implementation of AQ on well-performing meta-learners.