SolarWM-Wan2.2-5B / README.md
junchaoh-cs's picture
Use backbone-qualified repository name
14b42bf verified
|
Raw
History Blame Contribute Delete
2.98 kB
metadata
tags:
  - world-model
  - video-generation
  - camera-control
  - long-horizon-video
  - wan2.2
library_name: pytorch

SolarWM-Wan2.2-5B

Wan2.2 TI2V-5B checkpoints for SolarWM

Project Page · GitHub · SolarWM Collection · Dataset · Paper

SolarWM is an open foundation for building interactive video world models from data preparation through scalable training and long-horizon inference. This repository contains only the Wan2.2 TI2V-5B family, so users do not need to download the 14B, LTX-22B, or H3-33B releases.

Checkpoints

Directory Stage Frames Release contents
SolarWM-5B-base Base Base model and runtime assets
SolarWM-5B-bid-stage0p5-81f Stage0.5 81 Bidirectional camera-conditioned checkpoint
SolarWM-5B-bid-stage0p5-153f Stage0.5 153 Longer-context bidirectional checkpoint
SolarWM-5B-tf-stage1-81f Stage1 81 Teacher-forced AnyFlow checkpoint
SolarWM-5B-sgf-stage2-81f Stage2 81 Few-step causal SGF checkpoint

Download

hf download junchaoh-cs/SolarWM-Wan2.2-5B --local-dir ./SolarWM-models

To download one checkpoint only:

hf download junchaoh-cs/SolarWM-Wan2.2-5B \
  --include "SolarWM-5B-sgf-stage2-81f/*" \
  --local-dir ./SolarWM-models

Keep the directory layout unchanged. SolarWM configurations resolve the base and staged checkpoint paths from these release directories.

Inference

Install SolarWM from the source repository, activate the Wan runtime environment, and follow the Wan2.2 TI2V-5B guide. For Stage2 inference, set:

model.base_path=/path/to/SolarWM-models/SolarWM-5B-base
checkpoint.path=/path/to/SolarWM-models/SolarWM-5B-sgf-stage2-81f

License

SolarWM code is released under Apache-2.0. Model weights and bundled backbone assets remain subject to the license and notices included in each release directory; they are not relicensed by SolarWM.

Citation

@misc{huang2026solarwmopendatascalable,
      title={SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models},
      author={Junchao Huang and Guian Fang and Shengju Qian and Xianghao Kong and Zhuoran Zhao and Wei Huang and Yihua Du and Zixin Zhang and Justin Cui and Yuchao Gu and Yukang Chen and Xinting Hu and Tianyu He and Shaoshuai Shi and Zhuotao Tian and Xin Wang and Mike Zheng Shou and Li Jiang},
      year={2026},
      eprint={2609.02886},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2609.02886}
}