DRACO: Byzantine-resilient Distributed Training via Redundant Gradients
Primary LanguagePythonMIT LicenseMIT