# dusty-nv/jetson-inference

Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.

Repository: https://github.com/dusty-nv/jetson-inference
Canonical: https://ross.abutalabs.com/products/jetson-inference
Homepage: https://developer.nvidia.com/embedded/twodaystoademo
Language: C++
License: MIT
License Family: permissive
Topics: deep-learning, inference, computer-vision, embedded, image-recognition, object-detection, segmentation, jetson, jetson-tx1, jetson-tx2, jetson-xavier, nvidia, tensorrt, digits, caffe, video-analytics, robotics, machine-learning, jetson-nano, jetson-xavier-nx
Last push: 2025-10-16T17:34:47+00:00

## Health v2 (maintenance only)
Score: 44/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 47, release rhythm 8, longevity 100
- inputs: {"age_days": 3686, "days_push": 321, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8969, forks 3108 (observed 2026-08-28T04:10:26.551214+00:00)

## What it is
A C++/Python DNN inference library and tutorial guide ('Hello AI World') for deploying deep learning vision models on NVIDIA Jetson devices using TensorRT. It provides primitives for image classification, object detection, semantic segmentation, pose estimation, and action recognition, with live camera streaming, WebRTC webapps, and ROS/ROS2 support.

## Use cases
- run real-time object detection on a Jetson Nano camera feed
- classify images with pretrained ImageNet models on embedded GPU
- perform semantic segmentation onboard Jetson with TensorRT
- estimate human pose from live video on edge devices
- train custom models with PyTorch and deploy with TensorRT on Jetson
- build a robotics vision pipeline with ROS2 and Jetson
- create a WebRTC webapp streaming inference results from Jetson

## When to choose
- you are deploying deep learning inference on NVIDIA Jetson hardware
- you need optimized TensorRT-accelerated vision DNNs in C++ or Python
- you want a guided tutorial covering inference and transfer learning on edge devices
- you need camera streaming, ROS integration, or webapp demos on Jetson

## When to avoid
- you are targeting non-NVIDIA or non-Jetson hardware
- you need training-only workflows without inference deployment
- you want a general-purpose cloud or desktop inference framework
- you need LLM or generative AI inference rather than vision primitives

## Facets
- artifact type: library
- maturity: stable
- function: deep-learning, computer-vision, image-processing, machine-learning, llm-inference
- domain: deep-learning, computer-vision, embedded-systems, robotics, machine-learning
- platform: cpp, python, embedded
- tags: jetson, tensorrt, nvidia, object-detection, semantic-segmentation, pose-estimation, image-classification, edge-ai, ros, pytorch, tutorial, gpu, linux, docker

## Member repositories
- dusty-nv/jetson-inference (main) score 44

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:26.551214+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:24:26.016094+00:00, confidence not recorded.
  - readme: https://github.com/dusty-nv/jetson-inference (fetched 2026-08-28T04:10:26.551214+00:00, sha 5949694cd382)
  - homepage: https://developer.nvidia.com/embedded/twodaystoademo (fetched 2026-08-29T08:24:29.680390+00:00, sha edc4afc15ee2)
- Data as of 2026-08-30T08:39:29.467469+00:00.