How to set vscode launch.json (cuda-gdb) to debug cuda ops in the repository, such as quantize_q8_1 operator

[NNUCJ]

   I want to debug cuda kernel function using cuda_gdb , but not success. Compile the project using the following command
cmake -B build -DLLAMA_CUDA=ON -DCMAKE_BUILD_TYPE=Debug -DLLAMA_CUDA_FORCE_DMMV=ON -DLLAMA_CUDA_FORCE_MMQ=ON
   cmake log like:

-- CUDA found
-- Using CUDA architectures: 52;61;70;75
-- The CUDA compiler identification is NVIDIA 12.2.140
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda-12.2/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- CUDA host compiler is GNU 9.4.0