saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-14 18:12:05 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
onnxruntime/python/auto_examples/plot_train_convert_predict.html
Olivia Jain 2da5895b3a
updating python docs to python folder (#6897)
2021-03-09 10:21:38 -08:00

449 lines
No EOL
39 KiB
HTML
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Train, convert and predict with ONNX Runtime &#8212; ONNX Runtime 1.7.0 documentation</title>
<link rel="stylesheet" href="../static/pygments.css" type="text/css" />
<link rel="stylesheet" href="../static/alabaster.css" type="text/css" />
<link rel="stylesheet" type="text/css" href="../static/graphviz.css" />
<link rel="stylesheet" type="text/css" href="../static/gallery.css" />
<link rel="stylesheet" type="text/css" href="../static/gallery-binder.css" />
<link rel="stylesheet" type="text/css" href="../static/gallery-dataframe.css" />
<link rel="stylesheet" type="text/css" href="../static/gallery-rendered-html.css" />
<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>
<link rel="index" title="Index" href="../genindex.html" />
<link rel="search" title="Search" href="../search.html" />
<link rel="prev" title="Common errors with onnxruntime" href="plot_common_errors.html" />
<link rel="stylesheet" href="../static/custom.css" type="text/css" />
<meta name="viewport" content="width=device-width, initial-scale=0.9, maximum-scale=0.9" />
</head><body>
<div class="document">
<div class="documentwrapper">
<div class="bodywrapper">
<div class="body" role="main">
<div class="sphx-glr-download-link-note admonition note">
<p class="admonition-title">Note</p>
<p>Click <a class="reference internal" href="#sphx-glr-download-auto-examples-plot-train-convert-predict-py"><span class="std std-ref">here</span></a>
to download the full example code</p>
</div>
<div class="sphx-glr-example-title section" id="train-convert-and-predict-with-onnx-runtime">
<span id="l-logreg-example"></span><span id="sphx-glr-auto-examples-plot-train-convert-predict-py"></span><h1>Train, convert and predict with ONNX Runtime<a class="headerlink" href="#train-convert-and-predict-with-onnx-runtime" title="Permalink to this headline"></a></h1>
<p>This example demonstrates an end to end scenario
starting with the training of a machine learned model
to its use in its converted from.</p>
<div class="contents local topic" id="contents">
<ul class="simple">
<li><p><a class="reference internal" href="#train-a-logistic-regression" id="id1">Train a logistic regression</a></p></li>
<li><p><a class="reference internal" href="#conversion-to-onnx-format" id="id2">Conversion to ONNX format</a></p></li>
<li><p><a class="reference internal" href="#probabilities" id="id3">Probabilities</a></p></li>
<li><p><a class="reference internal" href="#benchmark-with-randomforest" id="id4">Benchmark with RandomForest</a></p></li>
</ul>
</div>
<div class="section" id="train-a-logistic-regression">
<h2><a class="toc-backref" href="#id1">Train a logistic regression</a><a class="headerlink" href="#train-a-logistic-regression" title="Permalink to this headline"></a></h2>
<p>The first step consists in retrieving the iris datset.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">sklearn.datasets</span> <span class="kn">import</span> <span class="n">load_iris</span>
<span class="n">iris</span> <span class="o">=</span> <span class="n">load_iris</span><span class="p">()</span>
<span class="n">X</span><span class="p">,</span> <span class="n">y</span> <span class="o">=</span> <span class="n">iris</span><span class="o">.</span><span class="n">data</span><span class="p">,</span> <span class="n">iris</span><span class="o">.</span><span class="n">target</span>
<span class="kn">from</span> <span class="nn">sklearn.model_selection</span> <span class="kn">import</span> <span class="n">train_test_split</span>
<span class="n">X_train</span><span class="p">,</span> <span class="n">X_test</span><span class="p">,</span> <span class="n">y_train</span><span class="p">,</span> <span class="n">y_test</span> <span class="o">=</span> <span class="n">train_test_split</span><span class="p">(</span><span class="n">X</span><span class="p">,</span> <span class="n">y</span><span class="p">)</span>
</pre></div>
</div>
<p>Then we fit a model.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">sklearn.linear_model</span> <span class="kn">import</span> <span class="n">LogisticRegression</span>
<span class="n">clr</span> <span class="o">=</span> <span class="n">LogisticRegression</span><span class="p">()</span>
<span class="n">clr</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>/opt/miniconda/lib/python3.7/site-packages/sklearn/linear_model/_logistic.py:765: ConvergenceWarning: lbfgs failed to converge (status=1):
STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.
Increase the number of iterations (max_iter) or scale the data as shown in:
https://scikit-learn.org/stable/modules/preprocessing.html
Please also refer to the documentation for alternative solver options:
https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression
extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)
LogisticRegression()
</pre></div>
</div>
<p>We compute the prediction on the test set
and we show the confusion matrix.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">sklearn.metrics</span> <span class="kn">import</span> <span class="n">confusion_matrix</span>
<span class="n">pred</span> <span class="o">=</span> <span class="n">clr</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">X_test</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">confusion_matrix</span><span class="p">(</span><span class="n">y_test</span><span class="p">,</span> <span class="n">pred</span><span class="p">))</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>[[10 0 0]
[ 0 14 0]
[ 0 1 13]]
</pre></div>
</div>
</div>
<div class="section" id="conversion-to-onnx-format">
<h2><a class="toc-backref" href="#id2">Conversion to ONNX format</a><a class="headerlink" href="#conversion-to-onnx-format" title="Permalink to this headline"></a></h2>
<p>We use module
<a class="reference external" href="https://github.com/onnx/sklearn-onnx">sklearn-onnx</a>
to convert the model into ONNX format.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">skl2onnx</span> <span class="kn">import</span> <span class="n">convert_sklearn</span>
<span class="kn">from</span> <span class="nn">skl2onnx.common.data_types</span> <span class="kn">import</span> <span class="n">FloatTensorType</span>
<span class="n">initial_type</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">&#39;float_input&#39;</span><span class="p">,</span> <span class="n">FloatTensorType</span><span class="p">([</span><span class="kc">None</span><span class="p">,</span> <span class="mi">4</span><span class="p">]))]</span>
<span class="n">onx</span> <span class="o">=</span> <span class="n">convert_sklearn</span><span class="p">(</span><span class="n">clr</span><span class="p">,</span> <span class="n">initial_types</span><span class="o">=</span><span class="n">initial_type</span><span class="p">)</span>
<span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="s2">&quot;logreg_iris.onnx&quot;</span><span class="p">,</span> <span class="s2">&quot;wb&quot;</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
<span class="n">f</span><span class="o">.</span><span class="n">write</span><span class="p">(</span><span class="n">onx</span><span class="o">.</span><span class="n">SerializeToString</span><span class="p">())</span>
</pre></div>
</div>
<p>We load the model with ONNX Runtime and look at
its input and output.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">import</span> <span class="nn">onnxruntime</span> <span class="k">as</span> <span class="nn">rt</span>
<span class="n">sess</span> <span class="o">=</span> <span class="n">rt</span><span class="o">.</span><span class="n">InferenceSession</span><span class="p">(</span><span class="s2">&quot;logreg_iris.onnx&quot;</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;input name=&#39;</span><span class="si">{}</span><span class="s2">&#39; and shape=</span><span class="si">{}</span><span class="s2">&quot;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
<span class="n">sess</span><span class="o">.</span><span class="n">get_inputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">sess</span><span class="o">.</span><span class="n">get_inputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">))</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;output name=&#39;</span><span class="si">{}</span><span class="s2">&#39; and shape=</span><span class="si">{}</span><span class="s2">&quot;</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
<span class="n">sess</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">name</span><span class="p">,</span> <span class="n">sess</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">shape</span><span class="p">))</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>input name=&#39;float_input&#39; and shape=[None, 4]
output name=&#39;output_label&#39; and shape=[None]
</pre></div>
</div>
<p>We compute the predictions.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">input_name</span> <span class="o">=</span> <span class="n">sess</span><span class="o">.</span><span class="n">get_inputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">name</span>
<span class="n">label_name</span> <span class="o">=</span> <span class="n">sess</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">name</span>
<span class="kn">import</span> <span class="nn">numpy</span>
<span class="n">pred_onx</span> <span class="o">=</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">label_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">X_test</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</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">confusion_matrix</span><span class="p">(</span><span class="n">pred</span><span class="p">,</span> <span class="n">pred_onx</span><span class="p">))</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>[[10 0 0]
[ 0 15 0]
[ 0 0 13]]
</pre></div>
</div>
<p>The prediction are perfectly identical.</p>
</div>
<div class="section" id="probabilities">
<h2><a class="toc-backref" href="#id3">Probabilities</a><a class="headerlink" href="#probabilities" title="Permalink to this headline"></a></h2>
<p>Probabilities are needed to compute other
relevant metrics such as the ROC Curve.
Let’s see how to get them first with
scikit-learn.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">prob_sklearn</span> <span class="o">=</span> <span class="n">clr</span><span class="o">.</span><span class="n">predict_proba</span><span class="p">(</span><span class="n">X_test</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="n">prob_sklearn</span><span class="p">[:</span><span class="mi">3</span><span class="p">])</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>[[4.15987711e-03 8.54898335e-01 1.40941788e-01]
[9.44228260e-01 5.57707615e-02 9.78002823e-07]
[5.17324282e-02 8.88396143e-01 5.98714288e-02]]
</pre></div>
</div>
<p>And then with ONNX Runtime.
The probabilies appear to be</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">prob_name</span> <span class="o">=</span> <span class="n">sess</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()[</span><span class="mi">1</span><span class="p">]</span><span class="o">.</span><span class="n">name</span>
<span class="n">prob_rt</span> <span class="o">=</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">prob_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">X_test</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</span><span class="p">)})[</span><span class="mi">0</span><span class="p">]</span>
<span class="kn">import</span> <span class="nn">pprint</span>
<span class="n">pprint</span><span class="o">.</span><span class="n">pprint</span><span class="p">(</span><span class="n">prob_rt</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="mi">3</span><span class="p">])</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>[{0: 0.0041598789393901825, 1: 0.8548984527587891, 2: 0.14094172418117523},
{0: 0.9442282915115356, 1: 0.055770788341760635, 2: 9.78002844931325e-07},
{0: 0.05173242464661598, 1: 0.888396143913269, 2: 0.05987146869301796}]
</pre></div>
</div>
<p>Let’s benchmark.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">timeit</span> <span class="kn">import</span> <span class="n">Timer</span>
<span class="k">def</span> <span class="nf">speed</span><span class="p">(</span><span class="n">inst</span><span class="p">,</span> <span class="n">number</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">repeat</span><span class="o">=</span><span class="mi">20</span><span class="p">):</span>
<span class="n">timer</span> <span class="o">=</span> <span class="n">Timer</span><span class="p">(</span><span class="n">inst</span><span class="p">,</span> <span class="nb">globals</span><span class="o">=</span><span class="nb">globals</span><span class="p">())</span>
<span class="n">raw</span> <span class="o">=</span> <span class="n">numpy</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">timer</span><span class="o">.</span><span class="n">repeat</span><span class="p">(</span><span class="n">repeat</span><span class="p">,</span> <span class="n">number</span><span class="o">=</span><span class="n">number</span><span class="p">))</span>
<span class="n">ave</span> <span class="o">=</span> <span class="n">raw</span><span class="o">.</span><span class="n">sum</span><span class="p">()</span> <span class="o">/</span> <span class="nb">len</span><span class="p">(</span><span class="n">raw</span><span class="p">)</span> <span class="o">/</span> <span class="n">number</span>
<span class="n">mi</span><span class="p">,</span> <span class="n">ma</span> <span class="o">=</span> <span class="n">raw</span><span class="o">.</span><span class="n">min</span><span class="p">()</span> <span class="o">/</span> <span class="n">number</span><span class="p">,</span> <span class="n">raw</span><span class="o">.</span><span class="n">max</span><span class="p">()</span> <span class="o">/</span> <span class="n">number</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Average </span><span class="si">%1.3g</span><span class="s2"> min=</span><span class="si">%1.3g</span><span class="s2"> max=</span><span class="si">%1.3g</span><span class="s2">&quot;</span> <span class="o">%</span> <span class="p">(</span><span class="n">ave</span><span class="p">,</span> <span class="n">mi</span><span class="p">,</span> <span class="n">ma</span><span class="p">))</span>
<span class="k">return</span> <span class="n">ave</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Execution time for clr.predict&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;clr.predict(X_test)&quot;</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Execution time for ONNX Runtime&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;sess.run([label_name], {input_name: X_test.astype(numpy.float32)})[0]&quot;</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time for clr.predict
Average 8.56e-05 min=7.72e-05 max=0.000101
Execution time for ONNX Runtime
Average 5.02e-05 min=4.61e-05 max=6.17e-05
5.0215695519000296e-05
</pre></div>
</div>
<p>Let’s benchmark a scenario similar to what a webservice
experiences: the model has to do one prediction at a time
as opposed to a batch of prediction.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="k">def</span> <span class="nf">loop</span><span class="p">(</span><span class="n">X_test</span><span class="p">,</span> <span class="n">fct</span><span class="p">,</span> <span class="n">n</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="n">nrow</span> <span class="o">=</span> <span class="n">X_test</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="k">if</span> <span class="n">n</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">n</span> <span class="o">=</span> <span class="n">nrow</span>
<span class="k">for</span> <span class="n">i</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">n</span><span class="p">):</span>
<span class="n">im</span> <span class="o">=</span> <span class="n">i</span> <span class="o">%</span> <span class="n">nrow</span>
<span class="n">fct</span><span class="p">(</span><span class="n">X_test</span><span class="p">[</span><span class="n">im</span><span class="p">:</span> <span class="n">im</span><span class="o">+</span><span class="mi">1</span><span class="p">])</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Execution time for clr.predict&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, clr.predict, 100)&quot;</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">sess_predict</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="k">return</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">label_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">x</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</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="s2">&quot;Execution time for sess_predict&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, sess_predict, 100)&quot;</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time for clr.predict
Average 0.00723 min=0.00694 max=0.00837
Execution time for sess_predict
Average 0.00156 min=0.00152 max=0.00166
0.0015644603711552918
</pre></div>
</div>
<p>Let’s do the same for the probabilities.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Execution time for predict_proba&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, clr.predict_proba, 100)&quot;</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">sess_predict_proba</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="k">return</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">prob_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">x</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</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="s2">&quot;Execution time for sess_predict_proba&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, sess_predict_proba, 100)&quot;</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time for predict_proba
Average 0.0108 min=0.0104 max=0.0115
Execution time for sess_predict_proba
Average 0.0017 min=0.00163 max=0.00188
0.0016972313076257703
</pre></div>
</div>
<p>This second comparison is better as
ONNX Runtime, in this experience,
computes the label and the probabilities
in every case.</p>
</div>
<div class="section" id="benchmark-with-randomforest">
<h2><a class="toc-backref" href="#id4">Benchmark with RandomForest</a><a class="headerlink" href="#benchmark-with-randomforest" title="Permalink to this headline"></a></h2>
<p>We first train and save a model in ONNX format.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">sklearn.ensemble</span> <span class="kn">import</span> <span class="n">RandomForestClassifier</span>
<span class="n">rf</span> <span class="o">=</span> <span class="n">RandomForestClassifier</span><span class="p">()</span>
<span class="n">rf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="n">initial_type</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">&#39;float_input&#39;</span><span class="p">,</span> <span class="n">FloatTensorType</span><span class="p">([</span><span class="mi">1</span><span class="p">,</span> <span class="mi">4</span><span class="p">]))]</span>
<span class="n">onx</span> <span class="o">=</span> <span class="n">convert_sklearn</span><span class="p">(</span><span class="n">rf</span><span class="p">,</span> <span class="n">initial_types</span><span class="o">=</span><span class="n">initial_type</span><span class="p">)</span>
<span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="s2">&quot;rf_iris.onnx&quot;</span><span class="p">,</span> <span class="s2">&quot;wb&quot;</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
<span class="n">f</span><span class="o">.</span><span class="n">write</span><span class="p">(</span><span class="n">onx</span><span class="o">.</span><span class="n">SerializeToString</span><span class="p">())</span>
</pre></div>
</div>
<p>We compare.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">sess</span> <span class="o">=</span> <span class="n">rt</span><span class="o">.</span><span class="n">InferenceSession</span><span class="p">(</span><span class="s2">&quot;rf_iris.onnx&quot;</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">sess_predict_proba_rf</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="k">return</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">prob_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">x</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</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="s2">&quot;Execution time for predict_proba&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, rf.predict_proba, 100)&quot;</span><span class="p">)</span>
<span class="nb">print</span><span class="p">(</span><span class="s2">&quot;Execution time for sess_predict_proba&quot;</span><span class="p">)</span>
<span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, sess_predict_proba_rf, 100)&quot;</span><span class="p">)</span>
</pre></div>
</div>
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>Execution time for predict_proba
Average 1.25 min=1.23 max=1.26
Execution time for sess_predict_proba
Average 0.00274 min=0.00245 max=0.00442
0.0027408220106735826
</pre></div>
</div>
<p>Let’s see with different number of trees.</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">measures</span> <span class="o">=</span> <span class="p">[]</span>
<span class="k">for</span> <span class="n">n_trees</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">5</span><span class="p">,</span> <span class="mi">51</span><span class="p">,</span> <span class="mi">5</span><span class="p">):</span>
<span class="nb">print</span><span class="p">(</span><span class="n">n_trees</span><span class="p">)</span>
<span class="n">rf</span> <span class="o">=</span> <span class="n">RandomForestClassifier</span><span class="p">(</span><span class="n">n_estimators</span><span class="o">=</span><span class="n">n_trees</span><span class="p">)</span>
<span class="n">rf</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">X_train</span><span class="p">,</span> <span class="n">y_train</span><span class="p">)</span>
<span class="n">initial_type</span> <span class="o">=</span> <span class="p">[(</span><span class="s1">&#39;float_input&#39;</span><span class="p">,</span> <span class="n">FloatTensorType</span><span class="p">([</span><span class="mi">1</span><span class="p">,</span> <span class="mi">4</span><span class="p">]))]</span>
<span class="n">onx</span> <span class="o">=</span> <span class="n">convert_sklearn</span><span class="p">(</span><span class="n">rf</span><span class="p">,</span> <span class="n">initial_types</span><span class="o">=</span><span class="n">initial_type</span><span class="p">)</span>
<span class="k">with</span> <span class="nb">open</span><span class="p">(</span><span class="s2">&quot;rf_iris_</span><span class="si">%d</span><span class="s2">.onnx&quot;</span> <span class="o">%</span> <span class="n">n_trees</span><span class="p">,</span> <span class="s2">&quot;wb&quot;</span><span class="p">)</span> <span class="k">as</span> <span class="n">f</span><span class="p">:</span>
<span class="n">f</span><span class="o">.</span><span class="n">write</span><span class="p">(</span><span class="n">onx</span><span class="o">.</span><span class="n">SerializeToString</span><span class="p">())</span>
<span class="n">sess</span> <span class="o">=</span> <span class="n">rt</span><span class="o">.</span><span class="n">InferenceSession</span><span class="p">(</span><span class="s2">&quot;rf_iris_</span><span class="si">%d</span><span class="s2">.onnx&quot;</span> <span class="o">%</span> <span class="n">n_trees</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">sess_predict_proba_loop</span><span class="p">(</span><span class="n">x</span><span class="p">):</span>
<span class="k">return</span> <span class="n">sess</span><span class="o">.</span><span class="n">run</span><span class="p">([</span><span class="n">prob_name</span><span class="p">],</span> <span class="p">{</span><span class="n">input_name</span><span class="p">:</span> <span class="n">x</span><span class="o">.</span><span class="n">astype</span><span class="p">(</span><span class="n">numpy</span><span class="o">.</span><span class="n">float32</span><span class="p">)})[</span><span class="mi">0</span><span class="p">]</span>
<span class="n">tsk</span> <span class="o">=</span> <span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, rf.predict_proba, 100)&quot;</span><span class="p">,</span> <span class="n">number</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">repeat</span><span class="o">=</span><span class="mi">5</span><span class="p">)</span>
<span class="n">trt</span> <span class="o">=</span> <span class="n">speed</span><span class="p">(</span><span class="s2">&quot;loop(X_test, sess_predict_proba_loop, 100)&quot;</span><span class="p">,</span> <span class="n">number</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">repeat</span><span class="o">=</span><span class="mi">5</span><span class="p">)</span>
<span class="n">measures</span><span class="o">.</span><span class="n">append</span><span class="p">({</span><span class="s1">&#39;n_trees&#39;</span><span class="p">:</span> <span class="n">n_trees</span><span class="p">,</span> <span class="s1">&#39;sklearn&#39;</span><span class="p">:</span> <span class="n">tsk</span><span class="p">,</span> <span class="s1">&#39;rt&#39;</span><span class="p">:</span> <span class="n">trt</span><span class="p">})</span>
<span class="kn">from</span> <span class="nn">pandas</span> <span class="kn">import</span> <span class="n">DataFrame</span>
<span class="n">df</span> <span class="o">=</span> <span class="n">DataFrame</span><span class="p">(</span><span class="n">measures</span><span class="p">)</span>
<span class="n">ax</span> <span class="o">=</span> <span class="n">df</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">x</span><span class="o">=</span><span class="s2">&quot;n_trees&quot;</span><span class="p">,</span> <span class="n">y</span><span class="o">=</span><span class="s2">&quot;sklearn&quot;</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="s2">&quot;scikit-learn&quot;</span><span class="p">,</span> <span class="n">c</span><span class="o">=</span><span class="s2">&quot;blue&quot;</span><span class="p">,</span> <span class="n">logy</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">df</span><span class="o">.</span><span class="n">plot</span><span class="p">(</span><span class="n">x</span><span class="o">=</span><span class="s2">&quot;n_trees&quot;</span><span class="p">,</span> <span class="n">y</span><span class="o">=</span><span class="s2">&quot;rt&quot;</span><span class="p">,</span> <span class="n">label</span><span class="o">=</span><span class="s2">&quot;onnxruntime&quot;</span><span class="p">,</span>
<span class="n">ax</span><span class="o">=</span><span class="n">ax</span><span class="p">,</span> <span class="n">c</span><span class="o">=</span><span class="s2">&quot;green&quot;</span><span class="p">,</span> <span class="n">logy</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="n">ax</span><span class="o">.</span><span class="n">set_xlabel</span><span class="p">(</span><span class="s2">&quot;Number of trees&quot;</span><span class="p">)</span>
<span class="n">ax</span><span class="o">.</span><span class="n">set_ylabel</span><span class="p">(</span><span class="s2">&quot;Prediction time (s)&quot;</span><span class="p">)</span>
<span class="n">ax</span><span class="o">.</span><span class="n">set_title</span><span class="p">(</span><span class="s2">&quot;Speed comparison between scikit-learn and ONNX Runtime</span><span class="se">\n</span><span class="s2">For a random forest on Iris dataset&quot;</span><span class="p">)</span>
<span class="n">ax</span><span class="o">.</span><span class="n">legend</span><span class="p">()</span>
</pre></div>
</div>
<img alt="Speed comparison between scikit-learn and ONNX Runtime For a random forest on Iris dataset" class="sphx-glr-single-img" src="../images/sphx_glr_plot_train_convert_predict_001.png" />
<p class="sphx-glr-script-out">Out:</p>
<div class="sphx-glr-script-out highlight-none notranslate"><div class="highlight"><pre><span></span>5
Average 0.11 min=0.107 max=0.118
Average 0.00169 min=0.00161 max=0.00177
10
Average 0.167 min=0.165 max=0.169
Average 0.0017 min=0.00165 max=0.00179
15
Average 0.228 min=0.227 max=0.231
Average 0.0017 min=0.00167 max=0.00172
20
Average 0.291 min=0.286 max=0.296
Average 0.00175 min=0.00173 max=0.00176
25
Average 0.346 min=0.342 max=0.35
Average 0.00174 min=0.00172 max=0.00181
30
Average 0.407 min=0.404 max=0.41
Average 0.0018 min=0.00174 max=0.0019
35
Average 0.463 min=0.459 max=0.467
Average 0.0018 min=0.00176 max=0.00187
40
Average 0.531 min=0.521 max=0.556
Average 0.0018 min=0.00179 max=0.00183
45
Average 0.582 min=0.577 max=0.597
Average 0.00187 min=0.00185 max=0.00189
50
Average 0.642 min=0.64 max=0.645
Average 0.00188 min=0.00185 max=0.00196
&lt;matplotlib.legend.Legend object at 0x7fca300dcda0&gt;
</pre></div>
</div>
<p class="sphx-glr-timing"><strong>Total running time of the script:</strong> ( 5 minutes 52.417 seconds)</p>
<div class="sphx-glr-footer class sphx-glr-footer-example docutils container" id="sphx-glr-download-auto-examples-plot-train-convert-predict-py">
<div class="sphx-glr-download sphx-glr-download-python docutils container">
<p><a class="reference download internal" download="" href="../downloads/c647c128e0cf2b3db04ce60b41ef1a14/plot_train_convert_predict.py"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Python</span> <span class="pre">source</span> <span class="pre">code:</span> <span class="pre">plot_train_convert_predict.py</span></code></a></p>
</div>
<div class="sphx-glr-download sphx-glr-download-jupyter docutils container">
<p><a class="reference download internal" download="" href="../downloads/1680115d3d937dfbb2d86adb705d9c5d/plot_train_convert_predict.ipynb"><code class="xref download docutils literal notranslate"><span class="pre">Download</span> <span class="pre">Jupyter</span> <span class="pre">notebook:</span> <span class="pre">plot_train_convert_predict.ipynb</span></code></a></p>
</div>
</div>
<p class="sphx-glr-signature"><a class="reference external" href="https://sphinx-gallery.github.io">Gallery generated by Sphinx-Gallery</a></p>
</div>
</div>
</div>
</div>
</div>
<div class="sphinxsidebar" role="navigation" aria-label="main navigation">
<div class="sphinxsidebarwrapper">
<p class="logo"><a href="../index.html">
<img class="logo" src="../static/ONNX_Runtime_icon.png" alt="Logo"/>
</a></p>
<h1 class="logo"><a href="../index.html">ONNX Runtime</a></h1>
<h3>Navigation</h3>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../tutorial.html">Tutorial</a></li>
<li class="toctree-l1"><a class="reference internal" href="../api_summary.html">API Summary</a></li>
<li class="toctree-l1"><a class="reference internal" href="index.html">Gallery of examples</a></li>
</ul>
<div class="relations">
<h3>Related Topics</h3>
<ul>
<li><a href="../index.html">Documentation overview</a><ul>
<li><a href="index.html">Gallery of examples</a><ul>
<li>Previous: <a href="plot_common_errors.html" title="previous chapter">Common errors with onnxruntime</a></li>
</ul></li>
</ul></li>
</ul>
</div>
<div id="searchbox" style="display: none" role="search">
<h3 id="searchlabel">Quick search</h3>
<div class="searchformwrapper">
<form class="search" action="../search.html" method="get">
<input type="text" name="q" aria-labelledby="searchlabel" />
<input type="submit" value="Go" />
</form>
</div>
</div>
<script>$('#searchbox').show(0);</script>
</div>
</div>
<div class="clearer"></div>
</div>
<div class="footer">
&copy;2018-2021, Microsoft.
|
Powered by <a href="http://sphinx-doc.org/">Sphinx 3.5.1</a>
&amp; <a href="https://github.com/bitprophet/alabaster">Alabaster 0.7.12</a>
|
<a href="../sources/auto_examples/plot_train_convert_predict.rst.txt"
rel="nofollow">Page source</a>
</div>
</body>
</html>
Reference in a new issue View git blame Copy permalink