---
title: abot-world-0
canonical_url: "https://www.modelscope.cn/studios/amap_cvlab/abot-world-0"
md_url: "https://www.modelscope.cn/studios/amap_cvlab/abot-world-0.md"
repository: amap_cvlab/abot-world-0
chinese_name: Abot-World
last_updated: 2026-08-03
sdk_type: docker
sdk_version: 6.17.3
downloads: 0
stars: 70
---
# abot-world-0
> abot-world-0 - amap_cvlab 在 ModelScope 创建的在线 Demo。ABot-World: Infinite Interactive World Rollout on a Single Desktop GPU
amap_cvlab/abot-world-0 是 ModelScope 魔搭社区上的在线可交互 Demo（创空间），基于 docker 6.17.3 构建，中文名为「Abot-World」。
- **Repository**: amap_cvlab/abot-world-0
- **SDK**: docker
- **SDK version**: 6.17.3
- **Downloads**: 0
- **Stars**: 70
- **Last updated**: 2026-08-03
Source: https://www.modelscope.cn/studios/amap_cvlab/abot-world-0
---
ABot-World: Infinite Interactive World Rollout on a Single Desktop GPU
> **ModelScope L20N Docker profile:** this directory verifies the deployed CC 12.0
> at container startup, builds SageAttention2 for the assigned GPU, enables the
> upstream FP8 per-token GEMM path, and adds fair single-GPU session leasing.
> See [DEPLOY_L20N.md](DEPLOY_L20N.md).
[](https://abot-world.amap.com)
[](https://reactor.inc/abot-world)
[](https://amap-cvlab.github.io/ABot-World/)
[](#)
[](https://github.com/amap-cvlab/ABot-World)
[](https://huggingface.co/acvlab/ABot-World-0-5B-LF)
[](https://huggingface.co/spaces/acvlab/abot-world-interactive)
[](https://modelscope.cn/models/amap_cvlab/ABot-World-0-5B-LF)
ABot-World Team
> **TL;DR:** ABot-World turns a single NVIDIA RTX 5090 desktop GPU into a real-time interactive world simulator, enabling infinite action-conditioned world rollout at 720P, 16 FPS, 1.2s latency, and 19GB GPU memory.
## 🚀 Key Highlights
* 🎮 **Action-Driven World Control:** Responds to user actions in real time, enabling continuous exploration instead of passive video playback.
* ⚡ **Real-Time Desktop Inference:** Runs at 720p and 16 FPS on a single NVIDIA RTX 5090 desktop GPU, with 1.2s latency and 19GB GPU memory.
* ♾️ **Infinite World Rollout:** Supports open-ended interactive world generation beyond fixed video-length limits.
* 🧠 **Open-Ended World Imagination:** Expands the world with new scenes and dynamics during rollout, avoiding scene lock-in, without prompt switching, by our *LongForcing* training.
## 📢 News
- 2026-07-13: ABot-World is now on [Reactor](https://reactor.inc/abot-world)!
- 2026-07-10: We have decided to open-source our `500-hour` video training dataset with accurate action annotations. Stay tuned—we plan to release it very soon.
- 2026-07-09: We release the causal student model `ABot-World-0-5B-LF`, inference code, our local gradio demo and online playground [ABot World Studio](https://abot-world.amap.com).
## 🛠️ Setup
> This installation was tested on: Ubuntu 22.04, CUDA 13.3, NVIDIA RTX 5090.
1. Clone the repository:
```bash
git clone https://github.com/amap-cvlab/ABot-World.git
cd ABot-World
```
2. Install dependencies using conda:
```bash
conda create -n aworld python=3.12 -y
conda activate aworld
pip install -r requirements.txt
```
3. Download checkpoints:
Download models using HuggingFace:
```bash
pip install -U "huggingface_hub"
hf download acvlab/ABot-World-0-5B-LF --local-dir ./checkpoints/ABot-World-0-5B-LF
```
Download models using ModelScope:
```bash
pip install -U "modelscope"
modelscope download "amap_cvlab/ABot-World-0-5B-LF" --local_dir ./checkpoints/ABot-World-0-5B-LF
```
After downloading, the project should have the following checkpoint structure:
```text
checkpoints/
└── ABot-World-0-5B-LF/
├── Wan2.2_VAE.pth
├── taew2_2.pth
├── models_t5_umt5-xxl-enc-bf16.pth
├── diffusion_pytorch_model.safetensors
└── google/umt5-xxl/
```
The checkpoint paths are configured in `configs/long_forcing_dmd.yaml` and
`configs/default_config.yaml`. The distilled generator weights are already
merged into `ABot-World-0-5B-LF/diffusion_pytorch_model.safetensors`.
## 🤗 Gradio Demo
```bash
bash web_client/run.sh
```
Select a GPU with:
```bash
CUDA_ID=0 bash web_client/run.sh
```
## License
This project is released under the Apache License 2.0. See `LICENSE`, `NOTICE`,
and `THIRD_PARTY_NOTICES.md` for copyright and third-party attribution details.
## 🤝 Acknowledgement
This project builds on and is inspired by the following open-source projects: [Causal Forcing](https://github.com/thu-ml/Causal-Forcing), [AngelSlim](https://github.com/tencent/AngelSlim), [LightX2V](https://github.com/ModelTC/LightX2V), [taehv](https://github.com/madebyollin/taehv), [Wan2.2](https://github.com/Wan-Video/Wan2.2), [Helios](https://github.com/PKU-YuanGroup/Helios), from which the optimized Triton RoPE and normalization kernels in `wan/modules/helios_kernels` are derived.
## 🗓️ Roadmap
- [x] Interactive Web Playground (ABot World Studio)
- [x] Inference Code Release
- [x] Local Gradio Demo Release
- [x] Causal Student Model Release
- [ ] Bidirectional Teacher Model Release
- [ ] 500-Hour Video Training Dataset with Accurate Action Annotations
- [ ] Technical Report (Arxiv)
## 📝 Citation
If you find our work helpful, please cite our paper:
```
@article{abot-world-0,
title={ABot-World-0: Infinite Interactive World Rollout on a Single Desktop GPU},
author={ABot-World Team},
year={2026}
}
```
## 🛰 Contact Us via WeChat Group
Feel free to contact us!
