# jkjung-avt/tensorrt_demos

TensorRT MODNet, YOLOv4, YOLOv3, SSD, MTCNN, and GoogLeNet

Repository: https://github.com/jkjung-avt/tensorrt_demos
Canonical: https://ross.abutalabs.com/products/tensorrt_demos
Homepage: https://jkjung-avt.github.io/
Language: Python
License: MIT
License Family: permissive
Topics: tensorrt, yolov4, yolov3, ssd-mobilenet, mtcnn, googlenet, modnet, object-detection, jetson
Last push: 2025-09-02T01:43:26+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 39, release rhythm 35, longevity 100
- inputs: {"age_days": 2663, "days_push": 366, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1792, forks 543 (observed 2026-08-28T04:05:37.164468+00:00)

## What it is
A collection of Python demo programs showing how to optimize and run deep learning models (YOLOv4, YOLOv3, SSD, MTCNN, GoogLeNet, MODNet) with NVIDIA TensorRT. It targets NVIDIA Jetson developer kits and x86_64 PCs with NVIDIA GPUs, providing optimized inference examples with benchmarked performance.

## Use cases
- run yolov4 object detection on jetson nano with tensorrt
- optimize deep learning models with tensorrt for jetson
- run face detection mtcnn on jetson xavier nx
- video matting with modnet on jetson
- benchmark ssd mobilenet inference speed on jetson
- run tensorrt inference on x86 gpu
- convert darknet or caffe models to tensorrt engines

## When to choose
- you need fast optimized inference of these specific models on NVIDIA Jetson hardware
- you want reference code for converting Caffe/TensorFlow/DarkNet/PyTorch models to TensorRT
- you are benchmarking object detection performance on embedded NVIDIA devices

## When to avoid
- you need a production-ready inference framework rather than demo examples
- you target non-NVIDIA hardware or CPUs
- you need models not covered by the demos (e.g., transformers or segmentation models)
- you want training code rather than inference

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, image-processing, computer-vision, gpu-computing, developer-tools
- domain: deep-learning, computer-vision, gpu-computing, embedded-systems, developer-tools
- platform: python, embedded
- tags: tensorrt, jetson, yolov4, yolov3, ssd-mobilenet, mtcnn, googlenet, modnet, object-detection, model-optimization, inference, linux, gpu

## Member repositories
- jkjung-avt/tensorrt_demos (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:37.164468+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 2026-08-30T03:22:58.411314+00:00, confidence not recorded.
  - readme: https://github.com/jkjung-avt/tensorrt_demos (fetched 2026-08-28T04:05:37.164468+00:00, sha 5f34ced22532)
  - homepage: https://jkjung-avt.github.io/ (fetched 2026-08-29T11:01:58.312334+00:00, sha 4701350547bf)
- Data as of 2026-08-30T08:39:29.467469+00:00.