GstInference/Benchmarks: Difference between revisions

From RidgeRun Developer Wiki
Line 450: Line 450:
</style>
</style>
<div id="Buttons_Model" style="margin: auto; width: 1300px; height: auto;">
<div id="Buttons_Model" style="margin: auto; width: 1300px; height: auto;">
<button class="button" id="show_inceptionv1">Show InceptionV1 </button>
<button class="button" id="show_inceptionv1_cpu">Show InceptionV1 </button>
<button class="button" id="show_inceptionv2">Show InceptionV2 </button>
<button class="button" id="show_inceptionv2_cpu">Show InceptionV2 </button>
<button class="button" id="show_inceptionv3">Show InceptionV3 </button>
<button class="button" id="show_inceptionv3_cpu">Show InceptionV3 </button>
<button class="button" id="show_inceptionv4">Show InceptionV4 </button>
<button class="button" id="show_inceptionv4_cpu">Show InceptionV4 </button>
<button class="button" id="show_tinyyolov2">Show TinyYoloV2 </button>
<button class="button" id="show_tinyyolov2">Show TinyYoloV2 </button>
<button class="button" id="show_tinyyolov3">Show TinyYoloV3 </button>
<button class="button" id="show_tinyyolov3">Show TinyYoloV3 </button>
Line 626: Line 626:


         /*Select the backend to filter the data to show in the chart*/
         /*Select the backend to filter the data to show in the chart*/
         var show_inceptionv1 = document.getElementById('show_inceptionv1');
         var show_inceptionv1_cpu = document.getElementById('show_inceptionv1_cpu');
         show_inceptionv1.onclick = function () {
         show_inceptionv1_cpu.onclick = function () {
           view.setColumns([0,1,2]);
           view.setColumns([0,1,2]);
           view.hideColumns([3,4,5,6,7,8,9,10,11,12]);
           view.hideColumns([3,4,5,6,7,8,9,10,11,12]);
           materialChart.draw(view, materialOptions);
           materialChart.draw(view, materialOptions);
         }
         }
         var show_inceptionv2 = document.getElementById("show_inceptionv2");
         var show_inceptionv2_cpu = document.getElementById("show_inceptionv2_cpu");
         show_inceptionv2.onclick = function() {
         show_inceptionv2_cpu.onclick = function() {
           view.setColumns([0,3,4]);
           view.setColumns([0,3,4]);
           view.hideColumns([1,2,5,6,7,8,9,10,11,12]);
           view.hideColumns([1,2,5,6,7,8,9,10,11,12]);
           materialChart.draw(view, materialOptions);
           materialChart.draw(view, materialOptions);
         }
         }
         var show_inceptionv3 = document.getElementById('show_inceptionv3');
         var show_inceptionv3_cpu = document.getElementById('show_inceptionv3_cpu');
         show_inceptionv3.onclick = function() {
         show_inceptionv3_cpu.onclick = function() {
           view.setColumns([0,5,6]);
           view.setColumns([0,5,6]);
           view.hideColumns([1,2,3,4,7,8,9,10,11,12]);
           view.hideColumns([1,2,3,4,7,8,9,10,11,12]);
           materialChart.draw(view, materialOptions);
           materialChart.draw(view, materialOptions);
         }
         }
         var show_inceptionv4 = document.getElementById("show_inceptionv4");
         var show_inceptionv4_cpu = document.getElementById("show_inceptionv4_cpu");
         show_inceptionv4.onclick = function() {
         show_inceptionv4_cpu.onclick = function() {
           view.setColumns([0,7,8]);
           view.setColumns([0,7,8]);
           view.hideColumns([1,2,3,4,5,6,9,10,11,12]);
           view.hideColumns([1,2,3,4,5,6,9,10,11,12]);

Revision as of 15:19, 27 March 2020




Previous: Example Applications/DispTec Index Next: Model Zoo




GstInference Benchmarks

The following benchmarks were run with a source video (1920x1080@60). With the following base GStreamer pipeline, and environment variables:

$ VIDEO_FILE='video.mp4'
$ MODEL_LOCATION='graph_inceptionv1_tensorflow.pb'
$ INPUT_LAYER='input'
$ OUTPUT_LAYER='InceptionV1/Logits/Predictions/Reshape_1'

The environment variables were changed accordingly with the used model (Inception V1,V2,V3 or V4)

GST_DEBUG=inception1:1 gst-launch-1.0 filesrc location=$VIDEO_FILE ! decodebin ! videoconvert ! videoscale ! queue ! net.sink_model inceptionv1 name=net model-location=$MODEL_LOCATION backend=tensorflow backend::input-layer=$INPUT_LAYER  backend::output-layer=$OUTPUT_LAYER net.src_model ! perf ! fakesink -v

The Desktop PC had the following specifications:

  • Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz
  • 8 GB RAM
  • Cedar [Radeon HD 5000/6000/7350/8350 Series]
  • Linux 4.15.0-54-generic x86_64 (Ubuntu 16.04)

The Jetson Xavier power modes used were 2 and 6 (more information: Supported Modes and Power Efficiency)

  • View current power mode:
$ sudo /usr/sbin/nvpmodel -q
  • Change current power mode:
sudo /usr/sbin/nvpmodel -m x

Where x is the power mode ID (e.g. 0, 1, 2, 3, 4, 5, 6).

Summary

Desktop PC CPU Library
Model Framerate CPU Usage
Inception V1 11.89 48
Inception V2 10.33 65
Inception V3 5.41 90
Inception V4 3.81 94
Jetson Xavier (15W) CPU Library GPU Library
Model Framerate CPU Usage Framerate CPU Usage
Inception V1 8.24 86 52.3 43
Inception V2 6.58 88 39.6 42
Inception V3 2.54 92 17.8 25
Inception V4 1.22 94 9.4 20
Jetson Xavier (30W) CPU Library GPU Library
Model Framerate CPU Usage Framerate CPU Usage
Inception V1 6.41 93 66.27 72
Inception V2 5.11 95 50.59 62
Inception V3 1.96 98 22.95 44
Inception V4 0.98 99 12.14 32

Framerate

thumb
thumb

CPU Usage

thumb
thumb

TensorFlow Lite Benchmarks

FPS measurement







CPU usage measurement







Test benchmark video

The following video was used to perform the benchmark tests.
To download the video press right click on the video and select 'Save video as' and save this in your computer.

Test benchmark video


Previous: Example Applications/DispTec Index Next: Model Zoo