File size: 3,322 Bytes
5d3f316 4c47db4 5d3f316 4c47db4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 | ---
license: apache-2.0
library_name: diffusers
pipeline_tag: robotics
tags:
- robotics
- world-model
- video-generation
- vision-action
- next-forcing
- robotwin
---
# Next Forcing — RoboTwin Post-trained Checkpoint
Post-trained checkpoint for **Next Forcing: Causal World Modeling with
Multi-Chunk Prediction**, evaluated on the RoboTwin 2.0 benchmark.
- 📄 Paper: https://arxiv.org/pdf/2606.11187
- 🌐 Project page: https://gangweix.github.io/next-forcing/
- 💻 Code: https://github.com/gangweix/next-forcing
## Model Description
Next Forcing addresses the myopic supervision problem in autoregressive video
world models: next-chunk denoising tends to learn local appearance shortcuts
instead of long-range dynamics, especially at high frame rates. Lightweight
**Multi-Chunk Prediction (MCP)** modules predict multiple future chunks through
a causal chain during training, providing dense temporal supervision to the
backbone.
This checkpoint is the RoboTwin post-trained model, built on top of the
[LingBot-VA](https://github.com/Robbyant/lingbot-va) codebase. It was post-trained
from [`gangweix/next-forcing-base`](https://huggingface.co/gangweix/next-forcing-base).
| | |
| --- | --- |
| Parameters | 6.7B (BF16) |
| Backbone layers | 30 |
| MCP depths | 3 (`mcp_blocks_per_depth=3`, collect layers `[3, 11, 19, 29]`) |
| Benchmark | RoboTwin 2.0, 50 bimanual manipulation tasks |
| Initialized from | [`next-forcing-base`](https://huggingface.co/gangweix/next-forcing-base) (5.1B) |
## Results
Average success rate on RoboTwin 2.0:
| Setting | LingBot-VA | **Next Forcing** |
| --- | ---: | ---: |
| Clean | 92.9 | **94.1** |
| Random | 91.5 | **93.5** |
## Repository Layout
```text
transformer/ Next Forcing backbone with MCP modules (enable_mcp=true)
vae/
text_encoder/
tokenizer/
```
## Usage
Clone the code and install the dependencies as described in the
[repository README](https://github.com/gangweix/next-forcing).
```bash
python -m pip install "huggingface_hub[cli]"
hf download gangweix/next-forcing-posttrain-robotwin \
--local-dir ./checkpoints/next-forcing-posttrain-robotwin
```
The evaluation code resolves model subfolders by path, so point
`NEXT_FORCING_MODEL_PATH` at the **local directory**, not at the Hub repository
id:
```bash
export NEXT_FORCING_MODEL_PATH=$PWD/checkpoints/next-forcing-posttrain-robotwin
export ROBOTWIN_ROOT=/path/to/your/RoboTwin
# Start the inference server on one GPU
CUDA_VISIBLE_DEVICES=0 bash evaluation/robotwin/launch_server.sh
# In another terminal, evaluate one task for 100 trials
bash evaluation/robotwin/launch_client.sh /path/to/eval_results adjust_bottle
```
RoboTwin evaluation requires a working RoboTwin 2.0 installation; see the
[official guide](https://robotwin-platform.github.io/doc/usage/robotwin-install.html).
## License
Released under the Apache License 2.0. Next Forcing is developed on top of the
LingBot-VA codebase; please retain the upstream attribution and license when
redistributing.
## Citation
```bibtex
@article{xu2026next,
title={Next Forcing: Causal World Modeling with Multi-Chunk Prediction},
author={Xu, Gangwei and Zhang, Qihang and Zhou, Jiaming and Zhu, Xing and Shen, Yujun and Yang, Xin and Xu, Yinghao},
journal={arXiv preprint arXiv:2606.11187},
year={2026}
}
```
|