---
license: cc0-1.0
---

# CoinRun Dataset

This is a large dataset of **50M video frames** and actions collected from the **CoinRun** environment (Cobbe et al., 2020) for training world models.   
The dataset enables reproducible, large-scale experiments in action-conditioned video prediction. It is meant to be used with [Jasmine](https://github.com/p-doom/jasmine), our JAX-based world modeling codebase.

---

### Dataset Summary
- **Environment:** CoinRun (Procgen Benchmark)  
- **Frames:** 50 million  
- **Resolution:** 64 × 64  
- **Format:** [`ArrayRecord`](https://github.com/google/array_record) (for fast I/O)  
- **Splits:** `train` / `val` / `test`  
- **License:** [CC0 1.0](https://creativecommons.org/publicdomain/zero/1.0/)  

---

## Usage
This dataset is part of the [Jasmine](https://github.com/p-doom/jasmine) repository release. Frames were collected from random agent rollouts. 

The ArrayRecord format enables efficient dataloading using Grain and is optimized for the [Jasmine dataloader](https://github.com/p-doom/jasmine/blob/main/jasmine/utils/dataloader.py).

You can download the dataset using the `huggingface-cli` tool.

```bash
huggingface-cli download --repo-type dataset p-doom/coinrun-dataset --local-dir <data_path>
```

To start a training run using Jasmine, simply pass the `train` and `val` split to the training script. 
```bash
python jasmine/baselines/maskgit/train_tokenizer_vqvae.py \
    --data_dir <data_path>/train \
    --val_data_dir <data_path>/val \
    ...
```
## Citation

If you use our CoinRun dataset, please cite our work:

```tex
@article{
    mahajan2025jasmine,
    title={Jasmine: A simple, performant and scalable JAX-based world modeling codebase},
    author={Mihir Mahajan and Alfred Nguyen and Franz Srambical and Stefan Bauer},
    journal = {p(doom) blog},
    year={2025},
    url={https://pdoom.org/jasmine.html},
    note = {https://pdoom.org/blog.html}
}