IAttention FP8#4209

Draft

narendasan wants to merge 3 commits into

mainfrom

narendasan/quantization_fixes

Commits on May 29, 2026

feat: Adding the scale normalize flag in IAttention layer using a pass to annotate target nodes

narendasan
authored and
zewenli98
committed
fix add a pass to insert softmax quantization node for decomposed attention

narendasan
authored and
zewenli98
committed
chore: address review comments

narendasan
authored and
zewenli98
committed