# CompreFace GPU

CompreFace GPU is a free, open-source face recognition service that uses a graphics processing unit (GPU) to quickly detect, verify, and identify people in photos and videos for self-hosted security, check-in. It is open source and self-hosted, installable on Docker Hub.

Canonical page: https://ourlocalhost.com/app/6501/compreface-gpu

## Description

CompreFace GPU is a free, open-source face recognition service that uses a graphics processing unit (GPU) to quickly detect, verify, and identify people in photos and videos for self-hosted security, check-in automation, or photo organization.

## Platform availability

Available on 0 of 10 platforms.

## Health score

27/100 (Low), based on 3 of 4 factors.
- Maintained: 0 (last commit 1103 days ago)
- Popular: 79 (8,429 GitHub stars)
- Easy to install: 15 (0 install platforms, install notes)

## Links

- GitHub: https://github.com/exadel-inc/CompreFace
- Docker Hub: https://hub.docker.com/r/exadel
- Official website: https://exadel.com/accelerator-showcase/compreface/

## Common questions

### Do I need Docker before installing CompreFace?

Yes. Install Docker and Docker Compose before starting the app, because the start command will fail and the app won't launch without them.

### How do I keep the CompreFace services running?

Use docker-compose up -d to start the main services in the background. If you skip this, the services won't run and the app can't be used.

### How do I enable GPU acceleration?

GPU support is optional for faster face processing on an NVIDIA GPU. Start the container with GPU access, such as --gpus all, or it can't use the graphics card and processing will be slower.

### How can I integrate CompreFace into other systems?

It exposes a REST API for face recognition, face verification, face detection, landmark detection, age recognition, and gender recognition.

### Can I move it from Proxmox to Unraid for GPU use?

The thread describes a user moving from Proxmox to Unraid to use a custom GPU build. They noted the Proxmox host was a laptop and they could not upgrade its GPU.

## Alternatives

- Pipl: Business solutions for decoding online identities while linking them to real-world records.
- compreface: CompreFace is a free and open-source, Docker-deployed face recognition service that provides web APIs for face detection, recognition, veri…

---
Source: Our Local Host (https://ourlocalhost.com/app/6501/compreface-gpu), the self-hosted and open-source app directory.
