Week1&2: Introduction & Setup JetsonNano

• Most of this workshop is available in YouTube. Fell free to watch by yourself

Hello AI World

Hello AI World can be run completely onboard your Jetson, including inferencing with TensorRT and transfer learning with PyTorch. The inference portion of Hello AI World - which includes coding your own image classification and object detection applications for Python or C++, and live camera demos - can be run on your Jetson in roughly two hours or less, while transfer learning is best left to leave running overnight.

System Setup (previously done)

• Setting up Jetson with JetPack
• Running the Docker Container

Camera Setup

• Instruction about How to connect CSI Camera to the Jetson-Nano can be followed

Please send a message to the professor as soon as you finished

• Testing your camera running. Most of the information is extracted from this link
  • To check CSI camera, you can run nvgstcapture-1.0, which will start capture and preview display it on the screen:

nano@jetson-nano:~$ nvgstcapture-1.0 
Encoder null, cannot set bitrate!
Encoder Profile = High
Supported resolutions in case of ARGUS Camera
  (2) : 640x480
  (3) : 1280x720
  (4) : 1920x1080
.....

• Check rotation
• This example command will rotate the image 180 degrees (vertical flip)

nano@jetson-nano:~$ nvgstcapture-1.0 --orientation 2
.....

Take a picture and save to disk

1. Connect CSI camera
2. Execute in a shell the command nvgstcapture-1.0 --automate --capture-auto
3. Open File with eog nvcamtest_XX.jpg

Capture a video and save to disk

1. Connect CSI camera
2. Execute in a shell the command nvgstcapture-1.0 --mode=2 --automate --capture-auto
3. Application will record 10 seconds of video
4. Play File recorded with totem nvcamtest_XX.mp4

Please send a message to the professor as soon as you finished

Setup Container

• There are several pre-configured containers to be able to use the Jetson-Nano board
• The most common are related to their use for artificial intelligence and machine learning. The most popular containers hosted on NVIDIA GPU Cloud (NGC) are the following Docker container images for machine learning on Jetson:
  • l4t-ml
  • l4t-pytorch
  • l4t-tensorflow

Running Docker Container

• The pre-built Docker container images for this lab are hosted on DockerHub
  • The containers use the l4t-pytorch base container, so support for transfer learning / re-training is already included

Inference instructions

• Follow the github repo instructions to install the docker image in the Jetson-Nano

nano@jetson-nano:~$ git clone --recursive https://github.com/dusty-nv/jetson-inference
Cloning into 'jetson-inference'...
remote: Enumerating objects: 20861, done.
....

Launching the Container

• It's recommended to use the script docker/run.sh script to run the container
  • docker/run.sh will automatically pull the correct container tag from DockerHub based on your currently-installed version of JetPack-L4T
  • IMPORTANT: if you are using CSI (Rpi Camera): --volume /tmp/argus_socket:/tmp/argus_socket

nano@jetson-nano:~$ cd jetson-inference/
nano@jetson-nano:~/jetson-inference$ docker/run.sh --volume /tmp/argus_socket:/tmp/argus_socket
reading L4T version from /etc/nv_tegra_release
L4T BSP Version:  L4T R32.6.1
[sudo] password for nano: 
size of data/networks:  79397 bytes
.....

• For reference, the following paths automatically get mounted from your host device into the container:
  • jetson-inference/data (stores the network models, serialized TensorRT engines, and test images)
  • jetson-inference/python/training/classification/data (stores classification training datasets)
  • jetson-inference/python/training/classification/models (stores classification models trained by PyTorch)
  • jetson-inference/python/training/detection/ssd/data (stores detection training datasets)
  • jetson-inference/python/training/detection/ssd/models (stores detection models trained by PyTorch)

Running applications

• Once the container is up and running, you can then run example programs from the tutorial like normal inside the container:

root@jetson-nano:/jetson-inference# cd build/aarch64/bin
root@jetson-nano:/jetson-inference/build/aarch64/bin# ./video-viewer
# (press Ctrl+D to exit the container)

Please send a message to the professor as soon as you finished

Inference of Image Classification

root@jetson-nano:/jetson-inference# cd build/aarch64/bin
root@jetson-nano:/jetson-inference/build/aarch64/bin# ./imagenet images/jellyfish.jpg images/test/jellyfish.jpg

• Note that imagenet app classifies the image jellyfish.jpg as a jellyfish and store the image solution in the path data/images/test with a confidence of 99.85%

Please send a message to the professor as soon as you finished

Inference of Object Detection

root@jetson-nano:/jetson-inference# cd build/aarch64/bin
root@jetson-nano:/jetson-inference/build/aarch64/bin# ./detectnet images/peds_0.jpg images/test/peds_0.jpg

• Note that detectnet app detects four persons with a confidence of 70.0%, 97.6%, 98.4% and 86.1% and store the image solution in path data/images/test

Please send a message to the professor as soon as you finished

Using other IA Models

• You can Download other models with the script download-models.sh

nano@jetson-nano:~$ cd jetson-inference/tools
nano@jetson-nano:~/jetson-inference$ ./download-models.sh

Please send a message to the professor as soon as you finished