Skip to main content
Version: 3.9.0

GPU Usage

Since face recognition requires a lot of processing power, the GPU acceleration for the Face SDK modules is now available for running deep learning algorithms.

You can use GPU acceleration on:

  • Windows x86 64-bit
  • Linux x86 64-bit
  • Android
  • Jetson (JetPack 4.3/4.4)

In this section, you'll find the information about the GPU acceleration for available Face SDK modules, how to enable this function, as well as information about timing characteristics for the Face SDK modules with CPU and GPU usage, possible errors when using GPU, and solutions.

Desktop [beta]#

Currently, the GPU acceleration is available for the following modules (single GPU mode only):

To run models on GPU, you need to edit the configuration file of one of the supported recognizers: set use_cuda to 1.


You can also use pre-built docker containers with CUDA support, such as nvidia/cuda:10.0-cudnn7-devel-ubuntu16.04 (please note that some licenses can be unavailable in this case).

The GPU acceleration is performed on one of the available GPUs (by default on the GPU with index 0). The GPU index can be changed as follows:

  • via the gpu_index parameter in the configuration file
  • via the CUDA_VISIBLE_DEVICES environment variable (see more info about CUDA Environment Variables)


  • System requirements:
    • JetPack 4.3 or 4.4*

* Tests were performed on the Jetson TX2 and Jetson NX modules.

The archive with the required libraries is jetson_jetpack_4.3_4.4.tar.xz. By default, it uses the build for jetpack 4.4. If you need a build for jetpack 4.3, move all files from the lib/jetpack-4.3 directory to the lib directory.

Timing characteristics#

The table below shows the speed measurements for template creation using CPU and GPU:


See the timing characteristics of GPU-based face detection in the Face detection section.

Note: the NVIDIA GeForce GTX 1080 Ti and Intel Core i7 were used for the speed test.


` cannot open shared object file: No such file or directory`You need to set the environment variable `LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/path/to/lib`, where `path/to/lib` is the path to the lib directory from the `face_sdk` build, or move this file to `../lib` relative to the application you're running (since the run-time search path is set in ``).
Slow initializationIncreasing the default JIT cache size: `export CUDA_CACHE_MAXSIZE=2147483647` (see JIT Caching)

Android [beta]#

Currently, the GPU acceleration is available for the following modules:

The GPU usage can be enabled/disabled via the use_mobile_gpu flag in the configuration files of the Capturer, Recognizer, VideoWorker objects (in the configuration file of the VideoWorker object, GPU is enabled for detectors). By default, mobile GPU support is enabled (the value is 1). To disable the GPU usage, change the use_mobile_gpu flag to 0.

Timing characteristics#

The table below shows the speed measurements for the Face SDK modules using CPU and GPU:


Note: Google Pixel 3 was used for the speed test.

Last updated on