Fix stale Int4Tensor assertions in gemma4_31b CUDA pipeline tests#20230

Gasoonjia merged 1 commit into

mainpytorch/executorch:mainfrom

g4-fix-int4-coalesced-testpytorch/executorch:g4-fix-int4-coalesced-testCopy head branch name to clipboard