fix: skip DQ->MatMulNBits fusion when weight/scale initializer is shared#28326
Merged
tianleiwu merged 2 commits intoMay 12, 2026
Merged
Microsoft GitHub Policy Service / license/cla
succeeded
May 2, 2026 in 0s
All CLA requirements met.
This check verifies that the author has agreed to a CLA with Microsoft.
Loading