A pallet model trained with SDG optimized for NVIDIA Jetson.

A project demonstrating how to train your own gesture recognition deep learning pipeline. We start with a pre-trained detection model, repurpose it for hand detection using Transfer Learning Toolkit 3.0, and use it together with the purpose-built gesture recognition model. Once trained, we deploy this model on NVIDIA® Jetson™ using Deepstream SDK.

This repository contains training script for SSD-Mobilenet model to be used on Jetson device

A notebook that demonstrates how to use the NVIDIA Intelligent Video Analytics suite to detect objects in real-time. We use Transfer Learning Toolkit to train a fast and accurate detector and DeepStream to run that detector on an NVIDIA Jetson edge device.

Counting pigs app using Jetson Nano with custom-trained YOLOv7, SORT tracking and complete instructions.

In recent years, UAV began to appear in all aspects of production and life of human society, and has been widely used in aerial photography, monitoring, security, disaster relief and other fields. For example, UAV tracking can be used for urban security, automatic cruise to find suspects and assist in intelligent urban security management.However, the practical application of UAV in various early scenes was mostly based on human remote control or intervention, and the degree of automation was not high. The degree to which UAVs can be automated is one of the decisive factors in whether they can play a bigger role in the future. With the increasing demand of UAV automation, target tracking based on computer vision has become one of the current research hotspots. Some companies in China and abroad, such as DJI, have successfully equipped target tracking on UAVs, but these technologies only exist in papers and descriptions, and the specific implementation has not been sorted out and opened source. Therefore, we plan to try to complete this project by ourselves and open source it on Github. Traditional visual tracking has many advantages, such as strong autonomy, wide measurement range and access to a large amount of environmental information, it also has many disadvantages.It requires a powerful hardware system. In order to obtain accurate navigation information, it needs to be equipped with a high-resolution camera and a powerful processor. From image data acquisition to processing, huge data operations are involved, which undoubtedly increases the cost of UAV tracking. Moreover, the reliability of traditional visual navigation and tracking is poor, and it is difficult for UAV to work in complex lighting and obstacle scenes. Therefore, we plan to use deep learning for target tracking in this project. We can train our own model through deep learning algorithm (we have not decided what network structure to use), then move the trained model to the embedded development board for operation, fix it on the UAV, read the image through the camera and process the data, so that it can recognize the objects to be recognized and tracked. In this project, we will use NVIDIA Jetson TX2 development board, install ROS in Linux system, establish communication with pixhawk, and conduct UAV flight control through PID algorithm.

Object detection inference with Roboflow Train models on NVIDIA Jetson devices.

How to create, train and quantize network, then integrate it into pre/post image processing and generate CUDA C++ code for targeting Jetson AGX Xavier

Use 600 pcs of Masked and No_Masked people, trained and inferenced on Jetson Nano with Yolov3-Tiny

Interfaced Hokuyo Lidar and Razor IMU to Jetson TK1 to develop algorithms for obstacle detection and localization to generate and train Path using Adaptive Monte Carlo Localization.

Autonomous UAV-Based Maritime SAR using onboard YOLOv8/v11 for real-time swimmer detection and precision life buoy deployment. Combines GPS/IMU navigation, temporal tracking, and edge-optimized deep learning (Jetson/RPi). Models trained on augmented SeaDronesSee for high recall and low-latency inference in open-sea rescue missions.

Training material covering a brief introduction to the Jetson Nano hardware, software stack, deep learning basics, TensorRT, TensorFlow-TensorRT integration, Deepstream and a full hands-on walkthrough

Drone-Based Real-Time Military Vehicle Detection using YOLOv8/v11 for rapid identification of armored vehicles and convoys in aerial footage. Employs manually piloted UAVs with onboard object detection, leveraging edge-optimized deep learning (Jetson/RPi). Models trained on augmented aerial datasets to enhance detection accuracy in defense surveill

This repository is your ultimate guide to getting started with the NVIDIA Jetson Orin Nano. It covers both hardware setup and software configuration. We will use YOLOv11 to perform object detection with our Jetson Orin Nano. Additionally, we will use OpenCV to perform face detection with a pre-trained Haar cascades.

A tutorial on how to train a YOLOv4 vehicle detector using Darknet and the RoundaboutTraffic dataset on a NVIDIA Jetson Nano.

This system detects vehicles, license plates, and brand logos using YOLO models via RTSP. I annotated 500+ images to help train a new model (best.pt) for plate character recognition, tested it using C++/Python DeepStream, and deployed it on Windows and Jetson Nano.

Comprehensive guide to train SOTA Yolov7 models on custom data then accelerate and deploy on Nvidia Jetson through Deepstream

Explore deep reinforcement learning for autonomous mobile robots with this repository. It features a custom environment for training, virtual model testing, and real-world deployment on Nvidia Jetracer Pro powered with Jetson Nano. Train and assess AI model in Both Virtually and Physically

Using Nvidia Jetson Nano Board and Pytorch to train the autonomous RC car for the competition.

End-to-end ROS 2 autonomous robotics pipeline featuring a custom-trained YOLOv8 vehicle detection model optimized with TensorRT. Implements real-time Adaptive Cruise Control (ACC) on NVIDIA Jetson hardware

Custom Object Detection. Train with PyTorch and Deploy it efficiently on the Edge devices using TensorRT Engine. The Edge Devices include Nvidia Jetson Nano, TX!, TX2, Xavier, AGX Xavier and AGX Orin.

A deep-learning model to detect strabismus (eye misalignment) from images using a trained neural network. It includes a dataset, preprocessing pipeline, training code and an h5 model with ready deployment on a Jetson Nano/Windows with potential input from an ESP32-CAM.

A small prototype of the self-driving car using a Convolutional Neural Network. This is my course project to implement an end-to-end method for training convolutional neural networks for the autonomous navigation of a mobile robot. The proposed navigation system shows object detection. The vehicle it is used can directly output the linear velocity of the robot from an input image in a single step. The trained model gives wheel velocities for navigation as outputs in real-time making it possible to be implanted on mobile robots such as robotic vacuum cleaners. The experimental results show an average linear velocity with a maximum turning angle is 30 degrees. The proposed model built in python 3.0 and experimental tests based on small scale car Quanser latest self-driving car (QCar) state-of-the-art product equipped with Jetson TX2 has been conducted to verify the effectiveness of the proposed network.

Trains MNIST on an NVIDIA Jetson using TensorFlow.

Train and inference custom dataset on Jetson using TensorRT

During the training process of custom images in jetson nano developer kit, if you face issues like insufficient memory and low processing speed , you can use Label image tool. Here we are doing the main training in an external Linux environment and only the trained model with less amount of losses is copied to the board along with labels.txt file.

Playing around with NVIDIA's pre-trained models in the Transfer Learning Toolkit on a Jetson.

ArtDetectiveAI is an AI model trained on a Jetson Nano that specializes in detecting whether art is AI created or human created.

Walktrough for the setup of a Toolchain to use NVIDIA Turing GPUs to train a model that is optimized by TensorRT for inference on the Jetson TX2

This project introduces a system to detect solar panels from aerial imagery with a YOLOv8-based custom-trained deep learning model. The trained model has been optimized and deployed on an NVIDIA Jetson Orin Nano for real-time inference on edge devices