FastASR是一个比较方便的SPEECH TO TEXT的AI库。开源。下面介绍下其在交叉编译到ARM target时候的交叉编译的cmake配置：

cmake_minimum_required(VERSION 3.10)project(FastASR)SET(CMAKE_C_COMPILER "/home/xxx/buildroot/output/platform_name/host/bin/aarch64-buildroot-linux-gnu-gcc.br_real")SET(CMAKE_CXX_COMPILER "/home/xxx/buildroot/output/platform_name/host/bin/aarch64-buildroot-linux-gnu-g++.br_real")SET(CMAKE_LINKER  "/home/xxx/buildroot/output/platform_name/host/bin/aarch64-buildroot-linux-gnu-ld")SET(CMAKE_AR "/home/xxx/buildroot/output/platform_name/host/bin/aarch64-buildroot-linux-gnu-ar")set(CMAKE_CXX_STANDARD 11)set(CMAKE_POSITION_INDEPENDENT_CODE ON)SET(CMAKE_SYSROOT "/home/xxx/buildroot/output/platform_name/host/aarch64-buildroot-linux-gnu/sysroot/")option(FASTASR_BUILD_PYTHON_MODULE "build python module, using FastASR in Python" OFF)if(FASTASR_BUILD_PYTHON_MODULE)add_subdirectory("./third_party/pybind11")
endif()add_subdirectory("./third_party/webrtc")add_subdirectory(lib)add_subdirectory(examples)

配置好CMAKE后，还需要在buildroot下编译fftw3和openblas， 这个可以使用buildroot的make menuconfig 配置或者直接在buildroot的config下添加并打开对应的package的宏定义。

在buildroot下编译出来了fftw和openblas的库之后，便可以在FastASR根目录下mkdir build, 然后cd build去 执行cmake .. 及make。