rai-toolbox 0.2.1

PyTorch-centric library for evaluating and enhancing the robustness of AI technologies

The rAI-toolbox is designed to enable methods for evaluating and enhancing both the robustness and the explainability of AI models in a way that is scalable and that composes naturally with other popular ML frameworks.

A key design principle of the rAI-toolbox is that it adheres strictly to the APIs specified by the PyTorch machine learning framework. For example, the rAI-toolbox frames adversarial training workflows solely in terms of the torch.nn.Optimizer and torch.nn.Module APIs. This makes it trivial to leverage other libraries and frameworks from the PyTorch ecosystem to bolster your responsible AI R&D. For instance, one can naturally leverage the rAI-toolbox together with PyTorch Lightning to perform distributed adversarial training.