Home

Awesome

Inference using UniAD

This repository is a fork of the UniAD, used to showcase how to implment a model to be evaluated according to the NeuroNCAP evaluation framework.

Changes

This repository differs from the original UniAD repository in the following ways:

How to use

  1. Download the weights:
mkdir checkpoints
wget "https://github.com/OpenDriveLab/UniAD/releases/download/v1.0.1/uniad_base_e2e.pth" -P checkpoints
wget https://github.com/OpenDriveLab/UniAD/releases/download/v1.0/motion_anchor_infos_mode6.pkl -P checkpoints
  1. Build the .sif file:
docker build -t uniad:latest -f docker/Dockerfile .
singularity build uniad.sif docker-daemon://uniad:latest

Links:

  1. Follow the instructions in the NeuroNCAP repository.

Original UniAD README.md