ihunch-escape

Thanks!

• https://github.com/Daniil-Osokin/lightweight-human-pose-estimation-3d-demo.pytorch
• https://github.com/elastic7327/django-gunicorn-nginx-docker
• https://github.com/cr4zyd3v/django-docker-ssl

Usage

from requests import post
r = post('https://api.ihunch.koyo.io/upload', files={'file': open('demo.jpg', 'rb')})
d = r.json()
if d['human']:
    print('ihunch:', d['pred'])

Docker

docker run --name ihunch \
    -it -d \
    --gpus device=0 \
    -e DJANGO_SECRET_KEY="generated key" \
    -v ~/docker/etc/letsencrypt:/etc/letsencrypt \
    -p 443:443 \
    koyokr/ihunch-escape

Requirements

• CUDA 11.2 and CuDNN 8 (GeForce RTX 3080)
• OpenCV 4.5.1
• Python 3.8
• PyTorch 1.8.1
• Detectron2 0.4
• XGBoost 1.4.1
• Django REST framework 3.12.4
• Gunicorn 20.1.0
• Nginx 1.18.0
• Supervisord 4.1.0