Add Gemma 4 E2B QNN (Snapdragon Hexagon NPU) recipe — WIP#432

Open

justinchuby wants to merge 6 commits into

mainfrom

gemma4-qnn-recipe

Commits on May 27, 2026

Add Gemma 4 E2B QNN (Snapdragon Hexagon NPU) recipe — WIP
justinchuby
committed
Gemma 4 E2B QNN recipe: compile all 4 components, drop ModelBuilder surgeries
justinchuby
committed
Gemma 4 E2B QNN: restore MatMulNBitsToQDQ — QNN can't run MatMulNBits
justinchuby
committed
Gemma 4 QNN README: document standard Attention vs GroupQueryAttention
justinchuby
committed
Gemma 4 QNN: address Copilot review — pin versions, align names, doc env
justinchuby
committed
Gemma 4 QNN: unpin requirements / install commands
justinchuby
committed