# NVIDIA-AI-IOT/trt_pose

Real-time pose estimation accelerated with NVIDIA TensorRT

Repository: https://github.com/NVIDIA-AI-IOT/trt_pose
Canonical: https://ross.abutalabs.com/products/trt_pose
Language: Python
License: MIT
License Family: permissive
Topics: human-pose-estimation, human-pose, tensorrt, pytorch, torch2trt, jetson, jetson-nano, jetson-xavier, real-time, live-demo, pretrained-models
Last push: 2022-08-12T00:45:06+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2592, "days_push": 1483, "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 1065, forks 300 (observed 2026-08-28T04:03:26.685852+00:00)

## What it is
trt_pose is a Python library from NVIDIA for real-time human pose estimation accelerated with TensorRT, targeting NVIDIA Jetson and other NVIDIA GPU platforms. It ships pre-trained models and training scripts for keypoint detection tasks in MSCOCO format.

## Use cases
- detect human body keypoints in real time from a camera
- run pose estimation on Jetson Nano or Xavier
- train a custom keypoint detection model on COCO-format data
- accelerate PyTorch pose models with TensorRT
- build gesture or activity recognition on top of body keypoints

## When to choose
- you need low-latency pose estimation on NVIDIA Jetson hardware
- you want pre-trained human pose models optimized with TensorRT
- your keypoint task data is in MSCOCO format

## When to avoid
- you need pose estimation on non-NVIDIA hardware or CPU-only environments
- you want a maintained project with recent updates - the last release was 2022
- you need full-framework training pipelines beyond keypoint detection

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing
- domain: computer-vision, machine-learning, deep-learning, gpu-computing, iot
- platform: python, embedded
- tags: pose-estimation, tensorrt, nvidia-jetson, keypoint-detection, pytorch, real-time-inference, pretrained-models, gpu, linux

## Member repositories
- NVIDIA-AI-IOT/trt_pose (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:26.685852+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 2026-08-30T06:55:27.915179+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA-AI-IOT/trt_pose (fetched 2026-08-28T04:03:26.685852+00:00, sha e6011914a7cd)
- Data as of 2026-08-30T08:39:29.467469+00:00.