Deploying Deep Learning

Linux Version

Ubuntu 18.04
L4T-R32.5.0

Build Deep Learning on Jeston

Reference : https://github.com/dusty-nv/jetson-inference/blob/L4T-R32.5.0/docs/building-repo-2.md

Build Jetson Inference

Here's a condensed form of the commands to build/install the project directly on your Jetson.

$ sudo apt-get update
$ sudo apt-get install git cmake libpython3-dev python3-numpy
$ git clone --recursive https://github.com/dusty-nv/jetson-inference
$ cd jetson-inference
$ mkdir build
$ cd build
$ cmake ../
$ make -j$(nproc)
$ sudo make install
$ sudo ldconfig

Classifying Images with ImageNet

Reference : https://github.com/dusty-nv/jetson-inference/blob/L4T-R32.5.0/docs/imagenet-console-2.md

Using the ImageNet Program on Jetson

After building the project, make sure your terminal is located in the aarch64/bin directory.

$ cd jetson-inference/build/aarch64/bin

Next, let's classify an example image with the imagenet program, using Python.These images will then be easily viewable from your host device in the jetson-inference/data/images/test directory.

$ ./imagenet.py images/strawberry_0.jpg images/test/strawberry.jpg

strawberry

Locating Objects with DetectNet

Reference : https://github.com/dusty-nv/jetson-inference/blob/L4T-R32.5.0/docs/detectnet-console-2.md

Detecting Objects from Images

Make sure your terminal is located in the aarch64/bin directory.

$ cd jetson-inference/build/aarch64/bin

Here are some examples of detecting pedestrians in images with the default SSD-Mobilenet-v2 model.

$ ./detectnet.py --network=ssd-mobilenet-v2 images/peds_0.jpg images/test/person.jpg

person

Processing Video Files

You can also process videos from disk. There are some test videos found on your Jetson under /usr/share/visionworks/sources/data .

$ ./detectnet.py /usr/share/visionworks/sources/data/pedestrians.mp4 images/test/pedestrians_identify.mp4

pedestrians