SANA-WM

SANA-WM is NVlabs/Sana's camera-controlled world model family. FlashDreams includes the chunk-causal streaming release, exposed through the cam2v-sana-wm-streaming application, and the full-sequence bidirectional release, exposed as a programmatic pipeline configuration.

Quick Start

Cam2V

uv sync --package flashdreams-sana-wm --inexact
uv run --no-sync flashdreams-run-v2 cam2v-sana-wm-streaming \
  --mode webrtc --host 0.0.0.0 --port 8089 -- --example-data

Demo Presets

Cam2V

Developer Details

• PyTorch: 2.9 or newer.
• Precision: BF16 by default. FP8 Stage-1/refiner inference is available on Hopper or newer GPUs (sm_90+), and FP4 is available on Blackwell (sm_100+). These upstream precision flags belong to SANA-WM_streaming.

Performance (Outdated)

The charts below compare steady-state generation latency per produced chunk for FlashDreams SANA-WM_streaming and the official SANA-WM_streaming implementation under matched settings. Warmup runs and the first decoded chunk are excluded from the headline metric. These GB300 latency runs show the official implementation faster than FlashDreams for BF16, FP8, and FP4.

In these charts, Official Impl means the pinned NVlabs/Sana upstream implementation measured by the FlashDreams benchmark harness under matched settings. It is not the SANA-WM 80-scene benchmark result published by the model authors.

All charts use the same demo image/prompt, w-80,dw-40,w-80,aw-40 action path, 241 requested frames, one discarded warmup run, and three measured runs. The benchmark runs recorded FlashDreams commit bd0816e and upstream commit 6298508.

Bidirectional variant

SANA-WM_bidirectional renders a complete clip in one pass from a first frame, prompt, and camera trajectory. It has no registered V2 application slug; use the lower-level pipeline configuration:

The bidirectional model is available as a pipeline configuration:

from sana_wm.config import PIPELINE_SANA_WM_BIDIRECTIONAL

pipeline = PIPELINE_SANA_WM_BIDIRECTIONAL.setup().to("cuda").eval()

This is the integration's lower-level pipeline configuration.

Bidirectional profiling benchmark

The BF16 chart below compares steady-state in-process generation latency per generated clip for FlashDreams SANA-WM_bidirectional and the official SANA-WM_bidirectional implementation under matched settings on one NVIDIA GB300 GPU. FlashDreams measured 34,182.39 ms per clip versus 56,932.83 ms for the official implementation.

In this chart, Official Impl means the pinned NVlabs/Sana upstream implementation measured by the FlashDreams benchmark harness under matched settings. It is not the SANA-WM 80-scene benchmark result published by the model authors.

Samples

Citation

If you use SANA-WM, please cite the original SANA work:

@misc{xie2024sana,
      title={SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformers},
      author={Enze Xie and Junsong Chen and Junyu Chen and Han Cai and Haotian Tang and Yujun Lin and Zhekai Zhang and Muyang Li and Ligeng Zhu and Yao Lu and Song Han},
      year={2024},
      eprint={2410.10629},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}