File size: 3,234 Bytes
f79f4b2 2a96a9e 72b2a93 f79f4b2 72b2a93 a7c84ed 2a96a9e 72b2a93 23131a3 72b2a93 2a96a9e a7c84ed 72b2a93 | 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 | ---
license: other
license_name: ltx-2-community-license-agreement
license_link: LICENSE
base_model: Lightricks/LTX-2.3
tags:
- video-generation
- lora
- camera-control
- world-model
- social-interaction
---
# HelloWorld V1
The social-interaction LoRA for **HelloWorld: Enabling Socially Interactive
Characters in Video World Models**.
📄 Paper: [arXiv:2608.05070](https://arxiv.org/abs/2608.05070) · [HF paper page](https://huggingface.co/papers/2608.05070)
- **Base model:** LTX-2.3-22B distilled ([`Lightricks/LTX-2.3`](https://huggingface.co/Lightricks/LTX-2.3))
- **File:** `helloworld_lora_v1.safetensors` (rank-32 LoRA)
- **What it does:** conditions LTX-2.3 on a camera-warp video so the generated
clip follows a user-supplied camera trajectory, while an F-key interaction
window makes the on-screen character respond toward the camera — turning to
the viewer, waving, nodding, or speaking a short greeting.
## Example videos
[`example_videos.zip`](example_videos.zip) (86 MB) contains the generated
results of the **seven official examples** shipped with the GitHub repo
(same recipes as [`assets/examples/`](https://github.com/AlayaLab/HelloWorld/tree/main/assets/examples),
previewed as GIFs in the repo README) — full-quality 1280×704 @ 24 fps MP4s
with the WASD / arrow / `F` keyboard overlay and the co-generated audio track
(the character speaks in `garden_pond`, `park_yoga`, `anime_figures`,
`moor_crows`, and `lakeside_hikers`; `bear_mascot` and `shore_skeleton` are
silent by design).
## Usage
Inference code, environment setup, and seven ready-to-reproduce examples:
**https://github.com/AlayaLab/HelloWorld**
```bash
git clone https://github.com/AlayaLab/HelloWorld
# put this LoRA at HelloWorld/checkpoints/helloworld_lora_v1.safetensors
cd HelloWorld/inference
INPUT=examples_batch.json GPU=0 bash run_batch.sh
```
## License
This LoRA is fine-tuned from **LTX-2.3** and trained on data synthesized by
LTX-2.3, which makes it a *Derivative of LTX-2* as defined in the
**LTX-2 Community License Agreement**. Accordingly, it is distributed
exclusively under that agreement — the complete license text is included in
this repository as [`LICENSE`](LICENSE). In particular:
- Your use of this model (and anything you build on it) is subject to the
use-based restrictions in **Attachment A** of the agreement.
- **Commercial entities with annual revenues of at least $10,000,000** must
obtain a separate paid commercial license from Lightricks — see
<https://ltx.io/model/licensing>.
- If you redistribute this model or further derivatives of it, you must do so
under the same agreement and include a complete copy of the license.
LTX-2 is © Lightricks Ltd. This repository is not affiliated with or endorsed
by Lightricks; all rights in the base model remain with the Licensor.
## Citation
```bibtex
@article{ouyang2026helloworld,
title = {HelloWorld: Enabling Socially Interactive Characters in Video World Models},
author = {Ouyang, Liangyang and Liu, Ruicong and Chu, Xuangeng and Zhang, Kaipeng and Sato, Yoichi},
journal = {arXiv preprint arXiv:2608.05070},
year = {2026}
}
```
## Contact
oyly@iis.u-tokyo.ac.jp / liangyang.ouyang@shanda.com
|