Repository for building and publishing Cosmos dependencies.
Install system dependencies:
export XDG_BIN_HOME="${XDG_BIN_HOME:-$HOME/.local/bin}"
curl -LsSf https://astral.sh/uv/install.sh | sh
source $XDG_BIN_HOME/envpushd "$XDG_BIN_HOME" && curl https://zyedidia.github.io/eget.sh | sh && popd
eget casey/just --to="$XDG_BIN_HOME"If uploading wheels, gh:
eget cli/cli --asset=.tar.gz --to="$XDG_BIN_HOME"
gh auth loginRun the docker container:
just docker-<cuda_version>Optionally, test the environment:
just build-dummyBuild a single package:
just build <package_name> <package_version> <python_version> <torch_version>Example:
just docker-cu128
just build-dummy
just build natten 0.21.1 3.12 2.9On the host, fix file permissions:
just fix-permissions- Upload wheels
just upload "build/**/*.whl"- Create and locally host the package index
just index-serve- Open a PR and merge to cosmos-dependencies.
uv version --bump=minor
gh release create --repo nvidia-cosmos/cosmos-dependencies v$(uv version --short)- cuda
- gcc
- flash-attn
We thrive on community collaboration! NVIDIA-Cosmos wouldn't be where it is without contributions from developers like you. Check out our Contributing Guide to get started, and share your feedback through issues.
Big thanks 🙏 to everyone helping us push the boundaries of open-source physical AI!
This repository is licensed under the Apache License 2.0.