---
license: agpl-3.0
library_name: pytorch
tags:
  - static2dynamic
  - diffusion
  - generative-modeling
  - biology
  - microscopy
  - image-to-video
  - trajectory-inference
pipeline_tag: image-to-video
---

# Static2Dynamic

Pretrained checkpoints for **Static2Dynamic**, a method for reconstructing videos of unobservable cellular, developmental, and disease processes from static, unpaired snapshots.

- [Code repository](https://github.com/biocompibens/static2dynamic)
- [Preprint](https://www.biorxiv.org/content/10.64898/2026.05.18.725860)

## Checkpoints

Each checkpoint subfolder contains:

- `net/`: diffusion model config and weights
- `video_time_encoder/`: time conditioning encoder config and weights
- `dynamic/`: scheduler configuration
- `my_conf/`: run configuration
- `train_samples.parquet`: training split samples
- `test_samples.parquet`: test split samples
- `pseudotime_predictions/`: pseudotime predictions 


## Download example

Download one checkpoint folder:

```sh
hf download thethomasboyer/Static2Dynamic --include 'NASH_steato/*' --local-dir ./Static2Dynamic_models
```