Issue/136 - 使用add_rms_norm融合算子#162

Open

gongchensu wants to merge 3 commits intoInfiniTensor:mainfrom

gongchensu:Issue/136

Commits on Dec 24, 2025

LM增加add_rms_norm算子的改动

zhuyue
authored and
gongchensu
committed
Issue/136 - Use fused add_rms_norm to avoid redundant add computation in LlamaDecoderLayer.
gongchensu
committed

Commits on Dec 25, 2025

Fuse residual add operations across decoder layers to eliminate redundant kernel launches.
gongchensu
committed