Get Started#

Welcome to FlashDreams! This page will guide you from a fresh checkout of the repository to a running model. It uses NVIDIA OmniDreams, the interactive driving world model, as the example; the model gallery lists the run command for every other model.

Install#

FlashDreams uses the uv Python package manager (installation instructions). With uv installed, clone the repository and synchronize the OmniDreams workspace:

git clone https://github.com/NVIDIA/flashdreams.git
cd flashdreams
uv sync --package flashdreams-omnidreams --extra interactive-drive

Most runs need a Hugging Face token. For OmniDreams, use a token with read access to nvidia/omni-dreams-models and nvidia/omni-dreams-scenes:

export HF_TOKEN=<your-hf-token>

For container, caching, and other environment details, see the project README and Troubleshooting.

Run your first model#

Launch the OmniDreams interactive driving demo. It runs the world model and streams the generated camera view to a browser:

uv run --package flashdreams-omnidreams interactive-drive --stream-mjpeg :8080

Then open http://<server-ip>:8080/ in a browser on the same network (use localhost on the same machine). The first launch spends several minutes loading checkpoints and compiling kernels; later launches reuse the cached assets.

On VRAM-constrained GPUs, add --offload-text-encoder to reduce peak VRAM usage by about 15 GB:

uv run --package flashdreams-omnidreams interactive-drive \
    --stream-mjpeg :8080 \
    --offload-text-encoder

See NVIDIA OmniDreams for the offload trade-offs, scripted generation, scene variants, WebRTC serving, and multi-GPU options.

Where to next#