Pinned Loading

1. Distributionally-Robust-Colored-MNIST Distributionally-Robust-Colored-MNIST Public

   A study of adversarial robustness using a custom Colored MNIST dataset. Implements distributionally robust optimization for training and offers tools for comparison with FGSM adversarial training a…

   Python

2. Nonlinear-Transformers-In-Context-PGD Nonlinear-Transformers-In-Context-PGD Public

   Codebase for experimenting with linear and nonlinear Transformers and their ability to implement preconditioned gradient descent (PGD) in-context.

   Python