siweilian
/
FunASR


			
							123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347
							<!DOCTYPE html>
<html class="writer-html5" lang="en" >
<head>
  <meta charset="utf-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>ONNXRuntime-python &mdash; FunASR  documentation</title><link rel="stylesheet" href="../../../_static/css/theme.css" type="text/css" />
    <link rel="stylesheet" href="../../../_static/pygments.css" type="text/css" />
  <!--[if lt IE 9]>
    <script src="../../../_static/js/html5shiv.min.js"></script>
  <![endif]-->
  <script id="documentation_options" data-url_root="../../../" src="../../../_static/documentation_options.js"></script>
        <script src="../../../_static/jquery.js"></script>
        <script src="../../../_static/underscore.js"></script>
        <script src="../../../_static/doctools.js"></script>
        <script src="../../../_static/language_data.js"></script>
        <script crossorigin="anonymous" integrity="sha256-Ae2Vz/4ePdIu6ZyI/5ZGsYnb+m0JlOmKPjt6XZ9JJkA=" src="https://cdnjs.cloudflare.com/ajax/libs/require.js/2.3.4/require.min.js"></script>
    <script src="../../../_static/js/theme.js"></script>
    <link rel="index" title="Index" href="../../../genindex.html" />
    <link rel="search" title="Search" href="../../../search.html" /> 
</head>

<body class="wy-body-for-nav"> 
  <div class="wy-grid-for-nav">
    <nav data-toggle="wy-nav-shift" class="wy-nav-side">
      <div class="wy-side-scroll">
        <div class="wy-side-nav-search" >

          
          <a href="../../../index.html" class="icon icon-home">
            FunASR
          </a>
<div role="search">
  <form id="rtd-search-form" class="wy-form" action="../../../search.html" method="get">
    <input type="text" name="q" placeholder="Search docs" aria-label="Search docs" />
    <input type="hidden" name="check_keywords" value="yes" />
    <input type="hidden" name="area" value="default" />
  </form>
</div>
        </div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">
              <p class="caption"><span class="caption-text">Installation</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../installation/installation.html">Installation</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../installation/docker.html">Docker</a></li>
</ul>
<p class="caption"><span class="caption-text">Quick Start</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../funasr/quick_start.html">Quick Start</a></li>
</ul>
<p class="caption"><span class="caption-text">Academic Egs</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../academic_recipe/asr_recipe.html">Speech Recognition</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../academic_recipe/punc_recipe.html">Punctuation Restoration</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../academic_recipe/vad_recipe.html">Voice Activity Detection</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../academic_recipe/sv_recipe.html">Speaker Verification</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../academic_recipe/sd_recipe.html">Speaker Diarization</a></li>
</ul>
<p class="caption"><span class="caption-text">ModelScope Egs</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../modelscope_pipeline/quick_start.html">Quick Start</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../egs_modelscope/asr/TEMPLATE/README.html">Speech Recognition</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../egs_modelscope/vad/TEMPLATE/README.html">Voice Activity Detection</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../egs_modelscope/punctuation/TEMPLATE/README.html">Punctuation Restoration</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../egs_modelscope/tp/TEMPLATE/README.html">Timestamp Prediction (FA)</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../modelscope_pipeline/sv_pipeline.html">Speaker Verification</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../modelscope_pipeline/sd_pipeline.html">Speaker Diarization</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../modelscope_pipeline/itn_pipeline.html">Inverse Text Normalization (ITN)</a></li>
</ul>
<p class="caption"><span class="caption-text">Model Zoo</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../model_zoo/modelscope_models.html">Pretrained Models Released on ModelScope</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../model_zoo/huggingface_models.html">Pretrained Models on Huggingface</a></li>
</ul>
<p class="caption"><span class="caption-text">Runtime and Service</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../readme.html">FunASR Runtime Roadmap</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../docs/SDK_tutorial_online.html">FunASR Realtime Transcribe Service</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../docs/SDK_tutorial.html">Highlights</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../docs/SDK_tutorial.html#funasr-offline-file-transcription-service">FunASR Offline File Transcription Service</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../html5/readme.html">Speech Recognition Service Html5 Client Access Interface</a></li>
</ul>
<p class="caption"><span class="caption-text">Benchmark and Leaderboard</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../benchmark/benchmark_pipeline_cer.html">Leaderboard IO</a></li>
</ul>
<p class="caption"><span class="caption-text">Funasr Library</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/build_task.html">Build custom tasks</a></li>
</ul>
<p class="caption"><span class="caption-text">Papers</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/papers.html">Papers</a></li>
</ul>
<p class="caption"><span class="caption-text">Application</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/application.html">Audio Cut</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/application.html#realtime-speech-recognition">Realtime Speech Recognition</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/application.html#audio-chat">Audio Chat</a></li>
</ul>
<p class="caption"><span class="caption-text">FQA</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../reference/FQA.html">FQA</a></li>
</ul>

        </div>
      </div>
    </nav>

    <section data-toggle="wy-nav-shift" class="wy-nav-content-wrap"><nav class="wy-nav-top" aria-label="Mobile navigation menu" >
          <i data-toggle="wy-nav-top" class="fa fa-bars"></i>
          <a href="../../../index.html">FunASR</a>
      </nav>

      <div class="wy-nav-content">
        <div class="rst-content">
          <div role="navigation" aria-label="Page navigation">
  <ul class="wy-breadcrumbs">
      <li><a href="../../../index.html" class="icon icon-home" aria-label="Home"></a></li>
      <li class="breadcrumb-item active">ONNXRuntime-python</li>
      <li class="wy-breadcrumbs-aside">
            <a href="../../../_sources/runtime/python/onnxruntime/README.md.txt" rel="nofollow"> View page source</a>
      </li>
  </ul>
  <hr/>
</div>
          <div role="main" class="document" itemscope="itemscope" itemtype="http://schema.org/Article">
           <div itemprop="articleBody">
             
  <div class="section" id="onnxruntime-python">
<h1>ONNXRuntime-python<a class="headerlink" href="#onnxruntime-python" title="Permalink to this headline"></a></h1>
<div class="section" id="install-funasr-onnx">
<h2>Install <code class="docutils literal notranslate"><span class="pre">funasr-onnx</span></code><a class="headerlink" href="#install-funasr-onnx" title="Permalink to this headline"></a></h2>
<p>install from pip</p>
<div class="highlight-shell notranslate"><div class="highlight"><pre><span></span>pip install -U funasr-onnx
<span class="c1"># For the users in China, you could install with the command:</span>
<span class="c1"># pip install -U funasr-onnx -i https://mirror.sjtu.edu.cn/pypi/web/simple</span>
<span class="c1"># If you want to export .onnx file, you should install modelscope and funasr</span>
pip install -U modelscope funasr
<span class="c1"># For the users in China, you could install with the command:</span>
<span class="c1"># pip install -U modelscope funasr -i https://mirror.sjtu.edu.cn/pypi/web/simple</span>
</pre></div>
</div>
<p>or install from source code</p>
<div class="highlight-shell notranslate"><div class="highlight"><pre><span></span>git clone https://github.com/alibaba/FunASR.git <span class="o">&amp;&amp;</span> <span class="nb">cd</span> FunASR
<span class="nb">cd</span> funasr/runtime/python/onnxruntime
pip install -e ./
<span class="c1"># For the users in China, you could install with the command:</span>
<span class="c1"># pip install -e ./ -i https://mirror.sjtu.edu.cn/pypi/web/simple</span>
</pre></div>
</div>
</div>
<div class="section" id="inference-with-runtime">
<h2>Inference with runtime<a class="headerlink" href="#inference-with-runtime" title="Permalink to this headline"></a></h2>
<div class="section" id="speech-recognition">
<h3>Speech Recognition<a class="headerlink" href="#speech-recognition" title="Permalink to this headline"></a></h3>
<div class="section" id="paraformer">
<h4>Paraformer<a class="headerlink" href="#paraformer" title="Permalink to this headline"></a></h4>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">funasr_onnx</span> <span class="kn">import</span> <span class="n">Paraformer</span>
<span class="kn">from</span> <span class="nn">pathlib</span> <span class="kn">import</span> <span class="n">Path</span>

<span class="n">model_dir</span> <span class="o">=</span> <span class="s2">&quot;damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch&quot;</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">Paraformer</span><span class="p">(</span><span class="n">model_dir</span><span class="p">,</span> <span class="n">batch_size</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">quantize</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>

<span class="n">wav_path</span> <span class="o">=</span> <span class="p">[</span><span class="s1">&#39;</span><span class="si">{}</span><span class="s1">/.cache/modelscope/hub/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch/example/asr_example.wav&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">Path</span><span class="o">.</span><span class="n">home</span><span class="p">())]</span>

<span class="n">result</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">wav_path</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">result</span><span class="p">)</span>
</pre></div>
</div>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">model_dir</span></code>: model_name in modelscope or local path downloaded from modelscope. If the local path is set, it should contain <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code>, <code class="docutils literal notranslate"><span class="pre">config.yaml</span></code>, <code class="docutils literal notranslate"><span class="pre">am.mvn</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">batch_size</span></code>: <code class="docutils literal notranslate"><span class="pre">1</span></code> (Default), the batch size duration inference</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">device_id</span></code>: <code class="docutils literal notranslate"><span class="pre">-1</span></code> (Default), infer on CPU. If you want to infer with GPU, set it to gpu_id (Please make sure that you have install the onnxruntime-gpu)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">quantize</span></code>: <code class="docutils literal notranslate"><span class="pre">False</span></code> (Default), load the model of <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code>. If set <code class="docutils literal notranslate"><span class="pre">True</span></code>, load the model of <code class="docutils literal notranslate"><span class="pre">model_quant.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">intra_op_num_threads</span></code>: <code class="docutils literal notranslate"><span class="pre">4</span></code> (Default), sets the number of threads used for intraop parallelism on CPU</p></li>
</ul>
<p>Input: wav formt file, support formats: <code class="docutils literal notranslate"><span class="pre">str,</span> <span class="pre">np.ndarray,</span> <span class="pre">List[str]</span></code></p>
<p>Output: <code class="docutils literal notranslate"><span class="pre">List[str]</span></code>: recognition result</p>
</div>
<div class="section" id="paraformer-online">
<h4>Paraformer-online<a class="headerlink" href="#paraformer-online" title="Permalink to this headline"></a></h4>
</div>
</div>
<div class="section" id="voice-activity-detection">
<h3>Voice Activity Detection<a class="headerlink" href="#voice-activity-detection" title="Permalink to this headline"></a></h3>
<div class="section" id="fsmn-vad">
<h4>FSMN-VAD<a class="headerlink" href="#fsmn-vad" title="Permalink to this headline"></a></h4>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">funasr_onnx</span> <span class="kn">import</span> <span class="n">Fsmn_vad</span>
<span class="kn">from</span> <span class="nn">pathlib</span> <span class="kn">import</span> <span class="n">Path</span>

<span class="n">model_dir</span> <span class="o">=</span> <span class="s2">&quot;damo/speech_fsmn_vad_zh-cn-16k-common-pytorch&quot;</span>
<span class="n">wav_path</span> <span class="o">=</span> <span class="s1">&#39;</span><span class="si">{}</span><span class="s1">/.cache/modelscope/hub/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/example/vad_example.wav&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">Path</span><span class="o">.</span><span class="n">home</span><span class="p">())</span>

<span class="n">model</span> <span class="o">=</span> <span class="n">Fsmn_vad</span><span class="p">(</span><span class="n">model_dir</span><span class="p">)</span>

<span class="n">result</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">wav_path</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">result</span><span class="p">)</span>
</pre></div>
</div>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">model_dir</span></code>: model_name in modelscope or local path downloaded from modelscope. If the local path is set, it should contain <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code>, <code class="docutils literal notranslate"><span class="pre">config.yaml</span></code>, <code class="docutils literal notranslate"><span class="pre">am.mvn</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">batch_size</span></code>: <code class="docutils literal notranslate"><span class="pre">1</span></code> (Default), the batch size duration inference</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">device_id</span></code>: <code class="docutils literal notranslate"><span class="pre">-1</span></code> (Default), infer on CPU. If you want to infer with GPU, set it to gpu_id (Please make sure that you have install the onnxruntime-gpu)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">quantize</span></code>: <code class="docutils literal notranslate"><span class="pre">False</span></code> (Default), load the model of <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code>. If set <code class="docutils literal notranslate"><span class="pre">True</span></code>, load the model of <code class="docutils literal notranslate"><span class="pre">model_quant.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">intra_op_num_threads</span></code>: <code class="docutils literal notranslate"><span class="pre">4</span></code> (Default), sets the number of threads used for intraop parallelism on CPU</p></li>
</ul>
<p>Input: wav formt file, support formats: <code class="docutils literal notranslate"><span class="pre">str,</span> <span class="pre">np.ndarray,</span> <span class="pre">List[str]</span></code></p>
<p>Output: <code class="docutils literal notranslate"><span class="pre">List[str]</span></code>: recognition result</p>
</div>
<div class="section" id="fsmn-vad-online">
<h4>FSMN-VAD-online<a class="headerlink" href="#fsmn-vad-online" title="Permalink to this headline"></a></h4>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">funasr_onnx</span> <span class="kn">import</span> <span class="n">Fsmn_vad_online</span>
<span class="kn">import</span> <span class="nn">soundfile</span>
<span class="kn">from</span> <span class="nn">pathlib</span> <span class="kn">import</span> <span class="n">Path</span>

<span class="n">model_dir</span> <span class="o">=</span> <span class="s2">&quot;damo/speech_fsmn_vad_zh-cn-16k-common-pytorch&quot;</span>
<span class="n">wav_path</span> <span class="o">=</span> <span class="s1">&#39;</span><span class="si">{}</span><span class="s1">/.cache/modelscope/hub/damo/speech_fsmn_vad_zh-cn-16k-common-pytorch/example/vad_example.wav&#39;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span><span class="n">Path</span><span class="o">.</span><span class="n">home</span><span class="p">())</span>

<span class="n">model</span> <span class="o">=</span> <span class="n">Fsmn_vad_online</span><span class="p">(</span><span class="n">model_dir</span><span class="p">)</span>


<span class="c1">##online vad</span>
<span class="n">speech</span><span class="p">,</span> <span class="n">sample_rate</span> <span class="o">=</span> <span class="n">soundfile</span><span class="o">.</span><span class="n">read</span><span class="p">(</span><span class="n">wav_path</span><span class="p">)</span>
<span class="n">speech_length</span> <span class="o">=</span> <span class="n">speech</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
<span class="c1">#</span>
<span class="n">sample_offset</span> <span class="o">=</span> <span class="mi">0</span>
<span class="n">step</span> <span class="o">=</span> <span class="mi">1600</span>
<span class="n">param_dict</span> <span class="o">=</span> <span class="p">{</span><span class="s1">&#39;in_cache&#39;</span><span class="p">:</span> <span class="p">[]}</span>
<span class="k">for</span> <span class="n">sample_offset</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">0</span><span class="p">,</span> <span class="n">speech_length</span><span class="p">,</span> <span class="nb">min</span><span class="p">(</span><span class="n">step</span><span class="p">,</span> <span class="n">speech_length</span> <span class="o">-</span> <span class="n">sample_offset</span><span class="p">)):</span>
    <span class="k">if</span> <span class="n">sample_offset</span> <span class="o">+</span> <span class="n">step</span> <span class="o">&gt;=</span> <span class="n">speech_length</span> <span class="o">-</span> <span class="mi">1</span><span class="p">:</span>
        <span class="n">step</span> <span class="o">=</span> <span class="n">speech_length</span> <span class="o">-</span> <span class="n">sample_offset</span>
        <span class="n">is_final</span> <span class="o">=</span> <span class="kc">True</span>
    <span class="k">else</span><span class="p">:</span>
        <span class="n">is_final</span> <span class="o">=</span> <span class="kc">False</span>
    <span class="n">param_dict</span><span class="p">[</span><span class="s1">&#39;is_final&#39;</span><span class="p">]</span> <span class="o">=</span> <span class="n">is_final</span>
    <span class="n">segments_result</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">audio_in</span><span class="o">=</span><span class="n">speech</span><span class="p">[</span><span class="n">sample_offset</span><span class="p">:</span> <span class="n">sample_offset</span> <span class="o">+</span> <span class="n">step</span><span class="p">],</span>
                            <span class="n">param_dict</span><span class="o">=</span><span class="n">param_dict</span><span class="p">)</span>
    <span class="k">if</span> <span class="n">segments_result</span><span class="p">:</span>
        <span class="nb">print</span><span class="p">(</span><span class="n">segments_result</span><span class="p">)</span>
</pre></div>
</div>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">model_dir</span></code>: model_name in modelscope or local path downloaded from modelscope. If the local path is set, it should contain <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code>, <code class="docutils literal notranslate"><span class="pre">config.yaml</span></code>, <code class="docutils literal notranslate"><span class="pre">am.mvn</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">batch_size</span></code>: <code class="docutils literal notranslate"><span class="pre">1</span></code> (Default), the batch size duration inference</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">device_id</span></code>: <code class="docutils literal notranslate"><span class="pre">-1</span></code> (Default), infer on CPU. If you want to infer with GPU, set it to gpu_id (Please make sure that you have install the onnxruntime-gpu)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">quantize</span></code>: <code class="docutils literal notranslate"><span class="pre">False</span></code> (Default), load the model of <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code>. If set <code class="docutils literal notranslate"><span class="pre">True</span></code>, load the model of <code class="docutils literal notranslate"><span class="pre">model_quant.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">intra_op_num_threads</span></code>: <code class="docutils literal notranslate"><span class="pre">4</span></code> (Default), sets the number of threads used for intraop parallelism on CPU</p></li>
</ul>
<p>Input: wav formt file, support formats: <code class="docutils literal notranslate"><span class="pre">str,</span> <span class="pre">np.ndarray,</span> <span class="pre">List[str]</span></code></p>
<p>Output: <code class="docutils literal notranslate"><span class="pre">List[str]</span></code>: recognition result</p>
</div>
</div>
<div class="section" id="punctuation-restoration">
<h3>Punctuation Restoration<a class="headerlink" href="#punctuation-restoration" title="Permalink to this headline"></a></h3>
<div class="section" id="ct-transformer">
<h4>CT-Transformer<a class="headerlink" href="#ct-transformer" title="Permalink to this headline"></a></h4>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">funasr_onnx</span> <span class="kn">import</span> <span class="n">CT_Transformer</span>

<span class="n">model_dir</span> <span class="o">=</span> <span class="s2">&quot;damo/punc_ct-transformer_zh-cn-common-vocab272727-pytorch&quot;</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">CT_Transformer</span><span class="p">(</span><span class="n">model_dir</span><span class="p">)</span>

<span class="n">text_in</span><span class="o">=</span><span class="s2">&quot;跨境河流是养育沿岸人民的生命之源长期以来为帮助下游地区防灾减灾中方技术人员在上游地区极为恶劣的自然条件下克服巨大困难甚至冒着生命危险向印方提供汛期水文资料处理紧急事件中方重视印方在跨境河流问题上的关切愿意进一步完善双方联合工作机制凡是中方能做的我们都会去做而且会做得更好我请印度朋友们放心中国在上游的任何开发利用都会经过科学规划和论证兼顾上下游的利益&quot;</span>
<span class="n">result</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">text_in</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">result</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
</pre></div>
</div>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">model_dir</span></code>: model_name in modelscope or local path downloaded from modelscope. If the local path is set, it should contain <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code>, <code class="docutils literal notranslate"><span class="pre">config.yaml</span></code>, <code class="docutils literal notranslate"><span class="pre">am.mvn</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">device_id</span></code>: <code class="docutils literal notranslate"><span class="pre">-1</span></code> (Default), infer on CPU. If you want to infer with GPU, set it to gpu_id (Please make sure that you have install the onnxruntime-gpu)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">quantize</span></code>: <code class="docutils literal notranslate"><span class="pre">False</span></code> (Default), load the model of <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code>. If set <code class="docutils literal notranslate"><span class="pre">True</span></code>, load the model of <code class="docutils literal notranslate"><span class="pre">model_quant.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">intra_op_num_threads</span></code>: <code class="docutils literal notranslate"><span class="pre">4</span></code> (Default), sets the number of threads used for intraop parallelism on CPU</p></li>
</ul>
<p>Input: <code class="docutils literal notranslate"><span class="pre">str</span></code>, raw text of asr result</p>
<p>Output: <code class="docutils literal notranslate"><span class="pre">List[str]</span></code>: recognition result</p>
</div>
<div class="section" id="ct-transformer-online">
<h4>CT-Transformer-online<a class="headerlink" href="#ct-transformer-online" title="Permalink to this headline"></a></h4>
<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">funasr_onnx</span> <span class="kn">import</span> <span class="n">CT_Transformer_VadRealtime</span>

<span class="n">model_dir</span> <span class="o">=</span> <span class="s2">&quot;damo/punc_ct-transformer_zh-cn-common-vad_realtime-vocab272727&quot;</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">CT_Transformer_VadRealtime</span><span class="p">(</span><span class="n">model_dir</span><span class="p">)</span>

<span class="n">text_in</span>  <span class="o">=</span> <span class="s2">&quot;跨境河流是养育沿岸|人民的生命之源长期以来为帮助下游地区防灾减灾中方技术人员|在上游地区极为恶劣的自然条件下克服巨大困难甚至冒着生命危险|向印方提供汛期水文资料处理紧急事件中方重视印方在跨境河流&gt;问题上的关切|愿意进一步完善双方联合工作机制|凡是|中方能做的我们|都会去做而且会做得更好我请印度朋友们放心中国在上游的|任何开发利用都会经过科学|规划和论证兼顾上下游的利益&quot;</span>

<span class="n">vads</span> <span class="o">=</span> <span class="n">text_in</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s2">&quot;|&quot;</span><span class="p">)</span>
<span class="n">rec_result_all</span><span class="o">=</span><span class="s2">&quot;&quot;</span>
<span class="n">param_dict</span> <span class="o">=</span> <span class="p">{</span><span class="s2">&quot;cache&quot;</span><span class="p">:</span> <span class="p">[]}</span>
<span class="k">for</span> <span class="n">vad</span> <span class="ow">in</span> <span class="n">vads</span><span class="p">:</span>
    <span class="n">result</span> <span class="o">=</span> <span class="n">model</span><span class="p">(</span><span class="n">vad</span><span class="p">,</span> <span class="n">param_dict</span><span class="o">=</span><span class="n">param_dict</span><span class="p">)</span>
    <span class="n">rec_result_all</span> <span class="o">+=</span> <span class="n">result</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>

<span class="nb">print</span><span class="p">(</span><span class="n">rec_result_all</span><span class="p">)</span>
</pre></div>
</div>
<ul class="simple">
<li><p><code class="docutils literal notranslate"><span class="pre">model_dir</span></code>: model_name in modelscope or local path downloaded from modelscope. If the local path is set, it should contain <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code>, <code class="docutils literal notranslate"><span class="pre">config.yaml</span></code>, <code class="docutils literal notranslate"><span class="pre">am.mvn</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">device_id</span></code>: <code class="docutils literal notranslate"><span class="pre">-1</span></code> (Default), infer on CPU. If you want to infer with GPU, set it to gpu_id (Please make sure that you have install the onnxruntime-gpu)</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">quantize</span></code>: <code class="docutils literal notranslate"><span class="pre">False</span></code> (Default), load the model of <code class="docutils literal notranslate"><span class="pre">model.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code>. If set <code class="docutils literal notranslate"><span class="pre">True</span></code>, load the model of <code class="docutils literal notranslate"><span class="pre">model_quant.onnx</span></code> in <code class="docutils literal notranslate"><span class="pre">model_dir</span></code></p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">intra_op_num_threads</span></code>: <code class="docutils literal notranslate"><span class="pre">4</span></code> (Default), sets the number of threads used for intraop parallelism on CPU</p></li>
</ul>
<p>Input: <code class="docutils literal notranslate"><span class="pre">str</span></code>, raw text of asr result</p>
<p>Output: <code class="docutils literal notranslate"><span class="pre">List[str]</span></code>: recognition result</p>
</div>
</div>
</div>
<div class="section" id="performance-benchmark">
<h2>Performance benchmark<a class="headerlink" href="#performance-benchmark" title="Permalink to this headline"></a></h2>
<p>Please ref to <a class="reference external" href="https://github.com/alibaba-damo-academy/FunASR/blob/main/runtime/docs/benchmark_onnx.md">benchmark</a></p>
</div>
<div class="section" id="acknowledge">
<h2>Acknowledge<a class="headerlink" href="#acknowledge" title="Permalink to this headline"></a></h2>
<ol class="simple">
<li><p>This project is maintained by <a class="reference external" href="https://github.com/alibaba-damo-academy/FunASR">FunASR community</a>.</p></li>
<li><p>We partially refer <a class="reference external" href="https://github.com/RapidAI/RapidASR">SWHL</a> for onnxruntime (only for paraformer model).</p></li>
</ol>
</div>
</div>


           </div>
          </div>
          <footer>

  <hr/>

  <div role="contentinfo">
    <p>&#169; Copyright 2022, Speech Lab, Alibaba Group.</p>
  </div>

  Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a
    <a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>
    provided by <a href="https://readthedocs.org">Read the Docs</a>.
   

</footer>
        </div>
      </div>
    </section>
  </div>
  <script>
      jQuery(function () {
          SphinxRtdTheme.Navigation.enable(true);
      });
  </script> 

</body>
</html>