NVIDIA/flashdreams

NVIDIA/flashdreams

Description: high-performance inference and serving library for interactive autoregressive video and world models

Language: Python

License: NOASSERTION

Stars: 273

Forks: 14

Open issues: 44

Created: 2026-04-23T09:03:25Z

Pushed: 2026-06-11T01:51:57Z

Default branch: main

Fork: no

Archived: no

README:

FlashDreams is a high-performance inference and serving library for interactive autoregressive video and world models. It began as the optimized runtime behind the [NVIDIA OmniDreams closed-loop demo for GTC 2026][omnidreams-blog] and has grown into a general platform for real-time world-model applications across gaming, autonomous vehicles, robotics, simulated or virtual environments, and more.

[omnidreams-blog]: https://research.nvidia.com/labs/sil/projects/omnidreams-blog/

https://github.com/user-attachments/assets/2b000ce9-effe-4cc9-a227-5b4619413e4d

System Requirements

• NVIDIA GPU with 80 GB VRAM or more (e.g. H100 80GB), see notes below.
• NVIDIA driver from the R580 series or newer (compatible with CUDA 13.x)
• CUDA 13.x (PyTorch 2.11.0+cu130 and the nvidia-*-cu13 libraries are

resolved by uv sync. A system CUDA toolkit is needed only for the developer extras and is included in nvidia/cuda:13.2.1-cudnn-devel-ubuntu24.04)

• Python >= 3.10
• PyTorch >= 2.11.0+cu130 (>= 2.9 for bare PyPI library install)
• Linux x86-64 or arm64
• 100 GB+ free storage space recommended for environment and model checkpoints.
• Docker with the

NVIDIA Container Toolkit (optional, only for the container workflow)

> Development and testing were performed on GPUs with 80 GB of VRAM or more. > Inference can fail (out-of-memory) on consumer and even enthusiast GPUs. > Per-model GPU and VRAM requirements are listed on each model page in > the model gallery.

Quickstart

The complete setup is in the installation guide. Assuming uv is installed, the shortest viable path is:

git clone https://github.com/NVIDIA/flashdreams.git
cd flashdreams
uv sync --extra runners
export HF_TOKEN=
uv run flashdreams-run --help

Note for developers/maintainers you would want to run uv sync --extra dev --extra runners instead.

Then launch your first model by following the quickstart guide. For example, the offline Self-Forcing T2V quickstart command is:

uv run --project integrations/self_forcing \
flashdreams-run self-forcing-wan2.1-t2v-1.3b \
--total-blocks 7

You can also install FlashDreams as a library from PyPI:

pip install flashdreams

Try the interactive driving demo

Drive a world model in real time with the OmniDreams interactive-drive demo. See the [interactive demo guide](https://nvidia.github.io/flashdreams/main/models/omnidreams.html#launch-the-interactive-demo).

Supported models

FlashDreams ships first-party integrations under [integrations/](integrations/). Each model has a dedicated docs page with runner slugs, multi-GPU commands, and (where available) profiling benchmarks.

| Model | Family | | --- | --- | | Self-Forcing | Streaming Wan2.1 T2V | | OmniDreams | HDMap-conditioned driving world model | | LingBot-World | Camera-controllable I2V world model | | Wan2.1 | Bidirectional T2V / I2V | | Causal-Forcing | Streaming Wan2.1 T2V / I2V | | Causal Wan2.2 | FastVideo Causal Wan 2.2 14B MoE T2V | | FlashVSR | Streaming video super-resolution | | Cosmos-Predict2.5 | Bidirectional T2V / I2V |

See the model gallery and the new method guide to add your own.

Developer guides

• Inference pipeline overview
• Config system
• Add a new method

For day-to-day development:

uv sync --extra dev --extra runners
uv run pre-commit run -a
uv run pytest -m "not manual"

See [DEV.md](DEV.md) for repository-specific workflow notes.

Contributing

For how to contribute, see [CONTRIBUTING.md](CONTRIBUTING.md). New integrations, bug reports, feature requests, performance tuning, and documentation edits are all welcome.

Use GitHub Issues to report defects or request improvements.

Join us on the NVIDIA Omniverse Discord to share your results and take part in technical discussion! Channel: `#flashdreams`

Security

To report a potential security vulnerability, follow the coordinated disclosure process in [SECURITY.md](SECURITY.md).

License

FlashDreams is released under the [Apache License 2.0](LICENSE). Third-party components and their licenses are listed in [THIRD-PARTY-NOTICES](THIRD-PARTY-NOTICES) and [NOTICE](NOTICE). The repository is REUSE-compliant; see [REUSE.toml](REUSE.toml) and [LICENSES/](LICENSES/).

Citation

If FlashDreams is useful in your research or product, please cite the project:

@misc{flashdreams2026,
title = {FlashDreams: High-performance inference and serving for
interactive autoregressive video and world models},
author = {{FlashDreams Contributors}},
year = {2026},
howpublished = {\url{https://github.com/NVIDIA/flashdreams}},
}

@misc{nvidia2026omnidreams,
title={OmniDreams:…