[BUILD] Fix narrowing conversion from double to unsigned int build issue.

[harrisonGPU]

When I built the TACO compiler, I encountered a build issue:

/home/harrison/Projects/taco/test/tests-tensor_types.cpp:48:30: error: narrowing conversion of ‘1.0e+0’ from ‘double’ to ‘unsigned int’ [-Wnarrowing]
   48 |   map<vector<int>,TypeParam> vals = {{{0}, 1.0}, {{2}, 2.0}};
      |                              ^~~~
/home/harrison/Projects/taco/test/tests-tensor_types.cpp: In instantiation of ‘void gtest_case_VectorTensorTest_::types<gtest_TypeParam_>::TestBody() [with gtest_TypeParam_ = short unsigned int]’:
/home/harrison/Projects/taco/test/tests-tensor_types.cpp:42:1:   required from here
17961 |   gtest_case_##TestCaseName##_
      |   ^~~~~~~~~~~
/home/harrison/Projects/taco/test/tests-tensor_types.cpp:48:30: error: narrowing conversion of ‘1.0e+0’ from ‘double’ to ‘short unsigned int’ [-Wnarrowing]
   48 |   map<vector<int>,TypeParam> vals = {{{0}, 1.0}, {{2}, 2.0}};
      |                              ^~~~
ninja: build stopped: subcommand failed.

This error indicates a narrowing conversion from double to unsigned int, causing the build to fail.

[harrisonGPU]

Ping @zhang677 , @weiya711