exadel-inc/CompreFace v1.2.0

CompreFace 1.2.0

on GitHub

Features

• Updated CUDA to version 11.8 and added support for the new generation of Nvidia GPUs, including the Ada Lovelace and Hopper microarchitectures.
• Added support for the Pose Plugin in the user interface.
• Implemented the ability to skip the detection step in the recognize faces endpoint.
• Introduced the option to send embeddings instead of the entire image for recognition.
• Included a "max_detect_size" option in the .env file.
• Implemented user password recovery functionality.
• Added the status page on CompreFace to start
• Updated result face sorting
• Implemented model loading during container startup.

Fixes

• Performance optimization and memory leak fixes
• Other fixes

Migration from 1.1.0 version

• No special migration is needed

How to install:

1. Download CompreFace_1.2.0.zip archive or run:
   wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.2.0/CompreFace_1.2.0.zip' && unzip tmp.zip && rm tmp.zip
2. To start CompreFace run:
   docker-compose up -d
3. Open in your browser: http://localhost:8000/login
4. To stop properly run:
   docker-compose stop
5. To start again run:
   docker-compose start

More detailed install guide(and Windows instructions) is on our readme page.