© 2026 Buobe

Towards Free Normalization: Fusing Normalization into GEMM and Attention Kernels

Published by

on •1 min read

Code available at: https://github.com/facebookresearch/ads_model_kernel_library/tree/main/multi_cta_norm_fusion and https://github.com/facebookresearch/ads_model_kernel_library/tree/main/gdpa_megakernel

Towards Free Normalization: Fusing Normalization into GEMM and Attention Kernels | Buobe