# Track 2 🔥

code: [github](https://github.com/ChenHongruixuan/BRIGHT?tab=readme-ov-file)

## ingest

data:
- [zenodo](https://zenodo.org/records/14619798) ⭐️ - manually downloaded, 10.2 GB, `dfc25_track2_trainval.zip`
- [huggingface](https://huggingface.co/datasets/Kullervo/BRIGHT) - API access.

`dfc25_track2_trainval.zip` -> decompressed -> uploaded as `dfc25_track2_trainval`.

🔥

digest https://github.com/ChenHongruixuan/BRIGHT 🔥

dev,  🔥

```bash
@fusion bright ingest \
  dryrun,train=<count>,upload,val=<count> \
  <object-name>
```

digest https://www.grss-ieee.org/technical-committees/image-analysis-and-data-fusion/?tab=data-fusion-contest  🔥