Zihan-Su / Self_Gradient_Forcing_Plus

# Self Gradient Forcing Plus
### Decoupling Gradient Flows for Autoregressive Video Generation
 Zihan Su 1,2* , Junhao Zhuang 2*† , Yaowei Li 2 , Siwen Lu 1 , Haoran Li 2 , Lingen Li 3 , Haoyu Wu 2 
 Weiyang Jin 2 , Songchun Zhang 2 , Haoyang Huang 2 , Chun Yuan 1† , Zeyue Xue 2 , Nan Duan 2 
 1 Tsinghua University   2 Joy Future Academy, JD   3 The Chinese University of Hong Kong
 * Equal contribution.   † Corresponding authors. 
⭐ If Self Gradient Forcing Plus is useful for your research, please consider starring this repository.
## 🔥 News
- **2026-10-20**: We release our blog Streaming Video Generation: Past and Future, tracing the evolution of autoregressive video generation. We hope it offers useful insights to the community.
- **2026-10-9**: We release the code for reproducing the gradient-conflict phenomena observed at the SGF and teacher-forcing initialization checkpoints. See the reproduction guide. In fact, beyond Wan2.1-1.3B, we have also observed this phenomenon in LTX-2.5.
- **2026-10-7**: Paper, model checkpoints, inference scripts, and training code are publicly released.
## 🎬 Video Comparisons
Left to right: **Self Forcing · Self Gradient Forcing · Self Gradient Forcing Plus**. More videos are available on our project page.
**By the train window**
https://github.com/user-attachments/assets/1fd94100-b832-4202-80ad-de5c8158f86d
**A tiny monster, a candle**
https://github.com/user-attachments/assets/af3fb848-2927-4b46-b9f6-8d62bd1dd31f
## 🧠 Method Overview
We uncover conflicting gradients between context writing and denoising in SGF. These gradients partially cancel on shared parameters, motivating separate optimization.
SGF+ separates context-writing and denoising parameters to resolve conflicting gradient updates, improving autoregressive video generation and enabling rollouts of up to 24 hours from only 5s training windows.
## 🛠️ Installation
The environment follows the SGF setup.
 [code] 
Run the following commands from the repository root.
## ⬇️ Download Weights
 [code] 
The script uses the Hugging Face CLI command `hf` by default. Set `HF_CLI=huggingface-cli` if your environment still uses the older command name.
It downloads:
- Wan base models to `wan_models/Wan2.1-T2V-1.3B` and `wan_models/Wan2.1-T2V-14B`.
- Causal-Forcing AR initialization checkpoints to `checkpoints/init/chunkwise/ar_diffusion.pt` and `checkpoints/init/framewise/ar_diffusion.pt`.
- Released SGF+ inference checkpoints to `hf_weights/chunkwise/model.pt` and `hf_weights/framewise/model.pt`.
- The training prompt list to `prompts/vidprom_filtered_extended.txt`.
Run `hf auth login` first if authentication is required.
## 🚀 Inference
The default prompt file is `prompts/test_prompt.txt` with 8 prompts. The launcher uses 8 GPUs when at least 8 GPUs are visible; otherwise it falls back to single-GPU serial inference. By default it generates `963` latent frames, which decode to about 240 seconds of video at 16 fps.
The inference script takes the release setting name