Before running AlpaSim, you need to set up your environment with the required dependencies and access credentials.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/NVlabs/alpasim/llms.txt
Use this file to discover all available pages before exploring further.
System requirements
AlpaSim requires a system with:- NVIDIA GPU with CUDA 12.6 or greater
- Docker and Docker Compose
- Python environment (managed via
uv) - Hugging Face account for data access
Dependencies
Install uv
Install For more installation options, see the uv documentation.
uv, a fast Python package installer and resolver:Install Docker
Install Docker and required plugins:
- Ubuntu
- Other platforms
Install CUDA
Install CUDA 12.6 or greater:Download from NVIDIA CUDA Downloads and follow the instructions for your platform.
Install NVIDIA Container Toolkit
Install the NVIDIA Container Toolkit to enable GPU support in Docker containers:See the NVIDIA Container Toolkit installation guide for detailed instructions.
Hugging Face access
AlpaSim downloads simulation artifacts from Hugging Face. You need to set up access:Create a Hugging Face account
If you don’t already have one, create a free account at huggingface.co.
Generate an access token
Create an access token with read access:
- Go to Hugging Face settings
- Click “New token”
- Select “Read” access
- Generate and copy the token
Dataset information
AlpaSim uses the PhysicalAI Autonomous Vehicles NuRec dataset:- Dataset location: Hugging Face dataset
- Format details: See the dataset README for information on the contents of artifacts used to define scenes
data/nre-artifacts/all-usdzs when you run simulations that require them.
Verify installation
To verify your setup is complete:Next steps
Quick start
Now that your environment is set up, run your first simulation