Skip to content

1. Setup

SLEAP uses deep neural networks to learn how to predict poses from data. Training these models is much faster when using a GPU1 for acceleration.

If you know you have a GPU on your machine or have a Mac with Apple Silicon, you can install SLEAP locally and follow along this tutorial.

sleap-nn neural-network backend

The SLEAP GUI can be installed and used independently of the sleap-nn backend for labeling. However, for this tutorial it is important that you have sleap-nn installed with the correct PyTorch and CUDA versions according to your machine (ex. CPU or GPU).

To check which PyTorch and CUDA versions you should have installed, see here.

Install SLEAP locally

See the main SLEAP installation instructions for detailed installation instructions.

If you have either a Linux or Windows laptop with a GPU, or a Mac with Apple Silicon, SLEAP will work natively with hardware acceleration.

Next up: Importing data

  1. Graphics processing unit. This is a hardware component that parallelizes computations across thousands of cores, making them particularly effective for the algorithms used to train deep neural networks. ↩