2025 ICML ICML 2025

COGNATE: Acceleration of Sparse Tensor Programs on Emerging Hardware using Transfer Learning