Get Started with LabGym
Install the toolkit and quantify user-defined behaviors on your own videos.
Installation
In order to successfully install LabGym, you will need to install Python 3.9 or 3.10 and add Python to PATH.
If you are on a Windows or Linux system with an NVIDIA GPU, you will need the following installed to enable GPU-accelerated processing:
- CUDA Toolkit 11.8
- cuDNN
We recommend that you follow the instructions below for your operating system to ensure that the installation goes smoothly. However, if your system configuration requires that you install these dependencies in a different way (e.g. you’re using a computing cluster that uses Lmod to load Python) feel free to go ahead and do that.
The simplest way to install LabGym is using pip. You may also use pipx or Anaconda to install LabGym in an isolated environment. For more information, please see the system-specific instructions below.
- Windows
- macOS
- Linux/WSL
First-Run Checklist
- Select video(s)
- Train or load a model
- Analyze → export results