FunASR/funasr/runtime/onnxruntime/readme.md

ONNXRuntime-cpp

Export the model

Install modelscope and funasr

pip3 install torch torchaudio
pip install -U modelscope
pip install -U funasr

Export onnx model

python -m funasr.export.export_model --model-name damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch --export-dir ./export --type onnx --quantize True

Building for Linux/Unix

Download onnxruntime

# download an appropriate onnxruntime from https://github.com/microsoft/onnxruntime/releases/tag/v1.14.0
# here we get a copy of onnxruntime for linux 64
wget https://github.com/microsoft/onnxruntime/releases/download/v1.14.0/onnxruntime-linux-x64-1.14.0.tgz
tar -zxvf onnxruntime-linux-x64-1.14.0.tgz

Install openblas

sudo apt-get install libopenblas-dev #ubuntu
# sudo yum -y install openblas-devel #centos

Build runtime

git clone https://github.com/alibaba-damo-academy/FunASR.git && cd funasr/runtime/onnxruntime
mkdir build && cd build
cmake  -DCMAKE_BUILD_TYPE=release .. -DONNXRUNTIME_DIR=/path/to/onnxruntime-linux-x64-1.14.0
make

Run the demo

funasr-onnx-offline

./funasr-onnx-offline     [--wav-scp <string>] [--wav-path <string>]
                          [--punc-quant <string>] [--punc-dir <string>]
                          [--vad-quant <string>] [--vad-dir <string>]
                          [--quantize <string>] --model-dir <string>
                          [--] [--version] [-h]
Where:
   --model-dir <string>
     (required)  the asr model path, which contains model.onnx, config.yaml, am.mvn
   --quantize <string>
     false (Default), load the model of model.onnx in model_dir. If set true, load the model of model_quant.onnx in model_dir

   --vad-dir <string>
     the vad model path, which contains model.onnx, vad.yaml, vad.mvn
   --vad-quant <string>
     false (Default), load the model of model.onnx in vad_dir. If set true, load the model of model_quant.onnx in vad_dir

   --punc-dir <string>
     the punc model path, which contains model.onnx, punc.yaml
   --punc-quant <string>
     false (Default), load the model of model.onnx in punc_dir. If set true, load the model of model_quant.onnx in punc_dir

   --wav-scp <string>
     wave scp path
   --wav-path <string>
     wave file path
  
   Required: --model-dir <string>
   If use vad, please add: --vad-dir <string>
   If use punc, please add: --punc-dir <string>

For example:
./funasr-onnx-offline \
    --model-dir    ./asrmodel/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch \
    --quantize  true \
    --vad-dir   ./asrmodel/speech_fsmn_vad_zh-cn-16k-common-pytorch \
    --punc-dir  ./asrmodel/punc_ct-transformer_zh-cn-common-vocab272727-pytorch \
    --wav-path    ./vad_example.wav

funasr-onnx-offline-vad

./funasr-onnx-offline-vad     [--wav-scp <string>] [--wav-path <string>]
                              [--quantize <string>] --model-dir <string>
                              [--] [--version] [-h]
Where:
   --model-dir <string>
     (required)  the vad model path, which contains model.onnx, vad.yaml, vad.mvn
   --quantize <string>
     false (Default), load the model of model.onnx in model_dir. If set true, load the model of model_quant.onnx in model_dir
   --wav-scp <string>
     wave scp path
   --wav-path <string>
     wave file path

   Required: --model-dir <string>

For example:
./funasr-onnx-offline-vad \
    --model-dir   ./asrmodel/speech_fsmn_vad_zh-cn-16k-common-pytorch \
    --wav-path    ./vad_example.wav

funasr-onnx-offline-punc

./funasr-onnx-offline-punc    [--txt-path <string>] [--quantize <string>]
                               --model-dir <string> [--] [--version] [-h]
Where:
   --model-dir <string>
     (required)  the punc model path, which contains model.onnx, punc.yaml
   --quantize <string>
     false (Default), load the model of model.onnx in model_dir. If set true, load the model of model_quant.onnx in model_dir
   --txt-path <string>
     txt file path, one sentence per line

   Required: --model-dir <string>

For example:
./funasr-onnx-offline-punc \
    --model-dir  ./asrmodel/punc_ct-transformer_zh-cn-common-vocab272727-pytorch \
    --txt-path   ./punc_example.txt

Acknowledge

1. This project is maintained by FunASR community.
2. We acknowledge mayong for contributing the onnxruntime(cpp api).
3. We borrowed a lot of code from FastASR for audio frontend and text-postprocess.