Support building GPU-capable sherpa-onnx on Linux aarch64. (#1500) #440

	name: windows-arm64

	on:
	push:
	branches:
	- master
	tags:
	- 'v[0-9]+.[0-9]+.[0-9]+*'
	paths:
	- '.github/workflows/windows-arm64.yaml'
	- 'CMakeLists.txt'
	- 'cmake/**'
	- 'sherpa-onnx/csrc/*'
	pull_request:
	branches:
	- master
	paths:
	- '.github/workflows/windows-arm64.yaml'
	- 'CMakeLists.txt'
	- 'cmake/**'
	- 'sherpa-onnx/csrc/*'

	workflow_dispatch:

	concurrency:
	group: windows-arm64-${{ github.ref }}
	cancel-in-progress: true

	jobs:
	windows_arm64:
	name: shared-${{ matrix.shared_lib }} tts-${{ matrix.with_tts }}
	runs-on: ${{ matrix.os }}
	strategy:
	fail-fast: false
	matrix:
	os: [windows-latest]
	shared_lib: [ON]
	with_tts: [ON, OFF]

	steps:
	- uses: actions/checkout@v4
	with:
	fetch-depth: 0

	- name: Configure CMake
	shell: bash
	run: \|
	mkdir build
	cd build
	cmake \
	-A ARM64 \
	-DSHERPA_ONNX_ENABLE_TTS=${{ matrix.with_tts }} \
	-D CMAKE_BUILD_TYPE=Release \
	-D BUILD_SHARED_LIBS=${{ matrix.shared_lib }} \
	-D CMAKE_INSTALL_PREFIX=./install \
	-D BUILD_ESPEAK_NG_EXE=OFF \
	..

	- name: Build sherpa-onnx for windows
	shell: bash
	run: \|
	cd build
	cmake --build . --config Release -- -m:2
	cmake --build . --config Release --target install -- -m:2

	ls -lh ./bin/Release/sherpa-onnx.exe

Provide feedback