.. _new:

What's New in FindFace Server 12.240830.2
===========================================


.. rubric:: New Features:

* Added the ability to use :ref:`multiple attribute extraction models <multiple_extraction_models>` (e.g. ``face_liveness``) on the installations with a single ``extraction-api`` and ``liveness-api`` instances.

.. rubric:: Technical Changes:

* Updated TensorRT to v10.1.0. Now FindFace Server supports all modern video card architectures NVIDIA (Pascal, Volta, Turing, Ampere, Ada and Hopper). See :ref:`Requirements <requirements>`.
* The official FindFace Server :ref:`Distribution package <deploy_docker-compose>` is compiled and includes service images and Docker Compose files base configured for CPU- and GPU-acceleration. You just need to add and specify the models. Easily deploy and configure this to meet your business needs.

.. rubric:: Enhanced Algorithms:

* :ref:`Licensing <licensing-principles>` for video streams is implemented for specific detection objects (such as ``cameras_face``, ``camera_body``, ``cameras_car``, etc.). Recording (Video Recorder) licensing is now available for each video stream (camera) – ``cameras_video_recording``.