---
license: cc0-1.0
---

# Jasmine Diffusion Checkpoint

Pretrained **diffusion-based world model** from the [Jasmine](https://github.com/p-doom/jasmine) codebase.  
Trained on the **CoinRun** dataset for action-conditioned video generation using the **diffusion-forcing** objective (Chen et al., 2024).

---

### Model Details
- **Architecture:** ST-DiT (spatio-temporal diffusion transformer)  
- **Input:** 16-frame sequences (64×64) + latent actions  
- **Training Environment:** [CoinRun (Cobbe et al., 2020)](https://huggingface.co/datasets/p-doom/coinrun-dataset)  
- **Objective:** Diffusion forcing (x-prediction)