See https://github.com/k2-fsa/sherpa
This repo uses onnxruntime and does not depend on libtorch.
git clone --recursive --branch v1.12.1 https://github.com/Microsoft/onnxruntime
cd onnxruntime
./build.sh \
--config RelWithDebInfo \
--build_shared_lib \
--build_wheel \
--skip_tests \
--parallel 16
cd build/Linux/RelWithDebInfo
sudo make install
export LD_LIBRARY_PATH=/path/to/onnxruntime/build/Linux/RelWithDebInfo:$LD_LIBRARY_PATH
git clone https://github.com/k2-fsa/sherpa-onnx
cd sherpa-onnx
mkdir build
cd build
cmake -DONNXRUNTIME_ROOTDIR=/path/to/onnxruntime \
-DKALDI_NATIVE_IO_INSTALL_PREFIX=/path/to/kaldi_native_io ..
make
./bin/sherpa-onnx path/to/encoder.onnx \
path/to/decoder.onnx \
path/to/joiner.onnx \
path/to/joiner_encoder_proj.onnx \
path/to/joiner_decoder_proj.onnx \
path/to/tokens.txt \
greedy \
path/to/audio.wav