Text-to-Video
Diffusers
Safetensors
world-model
video-generation
image-to-video
interactive
distillation
Instructions to use AlayaLab/Evoke with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use AlayaLab/Evoke with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("AlayaLab/Evoke", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
Link the arXiv paper, point at AlayaLab/Evoke, fix the download command
Browse files
README.md
CHANGED
|
@@ -14,14 +14,15 @@ tags:
|
|
| 14 |
<h1 align="center">Alaya-EVOKE: From Linear-Scaling Supervision to Endless World</h1>
|
| 15 |
|
| 16 |
<p align="center">
|
| 17 |
-
<a href="https://github.com/
|
| 18 |
<a href="https://evoke-world.github.io/Evoke/"><img src="https://img.shields.io/badge/π_Project_Page-evoke--world.github.io-1a73e8.svg" alt="Project Page"></a>
|
| 19 |
-
<a href="
|
|
|
|
| 20 |
</p>
|
| 21 |
|
| 22 |
-
Model weights for **EVOKE**, a 3-step, CFG-free
|
| 23 |
-
**384 Γ 640 @ 24 fps** video and stays coherent over 30 s
|
| 24 |
-
GitHub repository β **this repository holds weights only.**
|
| 25 |
|
| 26 |
- β‘ **3 steps, zero CFG** β 1.5 s of video every 2.11 s on one H200, one forward per step.
|
| 27 |
- π **Endless, not windowed** β scene geometry lives in an external camera-indexed world state bank,
|
|
@@ -46,11 +47,11 @@ evoke/
|
|
| 46 |
## Usage
|
| 47 |
|
| 48 |
```bash
|
| 49 |
-
git clone https://github.com/
|
| 50 |
pip install -r requirements.txt
|
| 51 |
|
| 52 |
-
|
| 53 |
-
|
| 54 |
|
| 55 |
MODE=t2v NUM_CHUNKS=20 bash scripts/inference/infer_post_distill.sh
|
| 56 |
```
|
|
@@ -75,7 +76,10 @@ built on [LingBot-World](https://github.com/robbyant/lingbot-world).
|
|
| 75 |
|
| 76 |
```bibtex
|
| 77 |
@article{evoke2026,
|
| 78 |
-
title
|
| 79 |
-
|
|
|
|
|
|
|
|
|
|
| 80 |
}
|
| 81 |
```
|
|
|
|
| 14 |
<h1 align="center">Alaya-EVOKE: From Linear-Scaling Supervision to Endless World</h1>
|
| 15 |
|
| 16 |
<p align="center">
|
| 17 |
+
<a href="https://github.com/AlayaLab/Evoke"><img src="https://img.shields.io/badge/GitHub-AlayaLab/Evoke-181717.svg?logo=github" alt="GitHub"></a>
|
| 18 |
<a href="https://evoke-world.github.io/Evoke/"><img src="https://img.shields.io/badge/π_Project_Page-evoke--world.github.io-1a73e8.svg" alt="Project Page"></a>
|
| 19 |
+
<a href="https://arxiv.org/abs/2608.13546"><img src="https://img.shields.io/badge/arXiv-2608.13546-b31b1b.svg" alt="arXiv"></a>
|
| 20 |
+
<a href="https://huggingface.co/papers/2608.13546"><img src="https://img.shields.io/badge/π€_Paper-2608.13546-ffce1c.svg" alt="Paper page"></a>
|
| 21 |
</p>
|
| 22 |
|
| 23 |
+
Model weights for **EVOKE** ([paper](https://huggingface.co/papers/2608.13546)), a 3-step, CFG-free
|
| 24 |
+
interactive world model that generates **384 Γ 640 @ 24 fps** video and stays coherent over 30 s
|
| 25 |
+
rollouts. Code, docs and demos live in the GitHub repository β **this repository holds weights only.**
|
| 26 |
|
| 27 |
- β‘ **3 steps, zero CFG** β 1.5 s of video every 2.11 s on one H200, one forward per step.
|
| 28 |
- π **Endless, not windowed** β scene geometry lives in an external camera-indexed world state bank,
|
|
|
|
| 47 |
## Usage
|
| 48 |
|
| 49 |
```bash
|
| 50 |
+
git clone https://github.com/AlayaLab/Evoke && cd Evoke
|
| 51 |
pip install -r requirements.txt
|
| 52 |
|
| 53 |
+
hf download SII-YuanyangYin/Evoke --local-dir models
|
| 54 |
+
hf download pkqbajng/ViGeo --local-dir models/ViGeo1.1 # REQUIRED depth backend
|
| 55 |
|
| 56 |
MODE=t2v NUM_CHUNKS=20 bash scripts/inference/infer_post_distill.sh
|
| 57 |
```
|
|
|
|
| 76 |
|
| 77 |
```bibtex
|
| 78 |
@article{evoke2026,
|
| 79 |
+
title = {Alaya-EVOKE: From Linear-Scaling Supervision to Endless World},
|
| 80 |
+
author = {Yin, Yuanyang and Wang, Gongxuan and Zhan, Yifan and
|
| 81 |
+
Li, Chuanhao and Zhang, Kaipeng and Zhao, Feng},
|
| 82 |
+
journal = {arXiv preprint arXiv:2608.13546},
|
| 83 |
+
year = {2026},
|
| 84 |
}
|
| 85 |
```
|