---
license: cc-by-4.0
library_name: pytorch
tags:
- remote-sensing
- super-resolution
- denoising
- sentinel-2
- major-tom
pipeline_tag: image-to-image
---

# Pulchra-Small

Joint **denoising + 3× super-resolution** (noisy 30 m RGB → clean 10 m RGB).

- **Train:** [Major-TOM/Core-S2L2A](https://huggingface.co/datasets/Major-TOM/Core-S2L2A)
- **Eval:** [SatClean-Bench](https://huggingface.co/datasets/NoraResearchLab/SatClean-Bench)

## SatClean-Bench results

| Level | PSNR | SSIM | LPIPS |
|-------|------|------|-------|
| 1 | 29.523 | 0.7343 | 0.4860 |
| 2 | 28.300 | 0.7087 | 0.5260 |
| 3 | 24.749 | 0.6612 | 0.6528 |

Bicubic (level 2) PSNR: **20.773**

## Maintainer

**NORA Research Lab**

[Website](https://noraresearchlab.site) · [GitHub](https://github.com/Nora-Research-Lab) · [HF](https://huggingface.co/NoraResearchLab) · [LinkedIn](https://www.linkedin.com/company/nora-research-lab) · [X](https://x.com/noraresearchlab)