All the codes are tested in the following environment:
Linux (tested on Ubuntu 22.04)
CUDA 11.3 or higher
# Install a virtual environment.
conda create -n dreamer python=3.9
conda activate dreamer
Install nuplan-devkit
from source
cd WorldDreamer/third_party/nuplan-devkit
pip install -r requirements.txt
pip install -e .
Install Pytorch==1.10.2
and torchvision==0.11.3
pip install torch==1.10.2+cu113 torchvision==0.11.3+cu113 torchaudio==0.10.2+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
Install the source code for other third-party packages, with cd ${FOLDER}; pip install -e .
# Install third-party packages
third_party/
├── bevfusion -> based on db75150
├── diffusers -> based on v0.17.1 (afcca3916)
└── xformers -> minorly change 0.0.19 to install with pytorch1.10.2
Install the dependencies of WorldDreamer
cd ..
pip install -r requirements.txt