onnxruntime

69e70471 - [QNN-EP] Add CastLoneQFusion to transform Cast and QNode into Convert (#25667)

[QNN-EP] Add CastLoneQFusion to transform Cast and QNode into Convert (#25667) ### Description - Introduced `CastLoneQFusion` in QNNEP to fuse `Cast` followed by `QuantizeLinear` into a single `Convert` operation. - Added corresponding test cases for **UINT8-to-FLOAT** `Cast` combined with `QuantizeLinear`, covering various **QuantType** scenarios. ### Motivation and Context - To optimize the model by reducing unnecessary QDQ nodes, this fusion transformation has been implemented. --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>