Fix python detection in cuda.cmake (pytorch#130651)

If Python package has not been detected previously, call it here

This fixes regression introduced by pytorch#128801 that results in annoying, but harmless warning reported in pytorch#129777

Pull Request resolved: pytorch#130651
Approved by: https://github.com/Skylion007

cmake/public/cuda.cmake

@@ -134,6 +134,7 @@ endif()
 # find lbnvrtc.so
 set(CUDA_NVRTC_LIB "${CUDA_nvrtc_LIBRARY}" CACHE FILEPATH "")
 if(CUDA_NVRTC_LIB AND NOT CUDA_NVRTC_SHORTHASH)
+  find_package(Python COMPONENTS Interpreter)
   execute_process(
     COMMAND Python::Interpreter -c
     "import hashlib;hash=hashlib.sha256();hash.update(open('${CUDA_NVRTC_LIB}','rb').read());print(hash.hexdigest()[:8])"