CAFIN paper proposes GPU-native kernel fusion to speed second-order optimizers
A new arXiv paper, CAFIN, introduces CUDA-accelerated adaptive Fisher information matrix network optimization using GPU-native kernel fusion to improve second-order optimizer efficiency in deep neural network training.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-11-13 05:00 · arXiv cs.LG
CAFIN: CUDA-accelerated Adaptive Fisher Information Matrix Network Optimization via GPU-Native Kernel Fusion