Enable named data map extension in CUDA build (pytorch#14898)

Since pytorch#14861 we need to specify
`EXECUTORCH_BUILD_EXTENSION_NAMED_DATA_MAP=ON` along with
`EXECUTORCH_BUILD_EXTENSION_MODULE=ON`. Adding that to fix CUDA CI.

Author: larryliu0820

.github/workflows/cuda.yml

@@ -136,6 +136,7 @@ jobs:
               -DEXECUTORCH_BUILD_CUDA=ON \
               -DEXECUTORCH_BUILD_EXTENSION_TENSOR=ON \
               -DEXECUTORCH_BUILD_EXTENSION_MODULE=ON \
+              -DEXECUTORCH_BUILD_EXTENSION_NAMED_DATA_MAP=ON \
               -DEXECUTORCH_BUILD_TESTS=ON \
               -Bcmake-out .
         cmake --build cmake-out -j$(( $(nproc) - 1 )) --target voxtral_runner