Hugging Face's logo

Datasets:
zl2048
/
SC-Captioner-data

Modalities:
Text
Formats:
json
Size:
10K - 100K
ArXiv:
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files
xet
Community
1
SC-Captioner-data
File size: 553 Bytes 
514cee2
 
 
 
 
545a60a
 
514cee2
 
 
fafd5a4
 
1
2
3
4
5
6
7
8
9
10
11
12
 
---
license: mit
---

## Training and testing annotations for SC-Captioner.
[![arXiv](https://img.shields.io/badge/cs.CV-2508.06125-b31b1b?logo=arxiv&logoColor=red)](https://arxiv.org/abs/2508.06125)
[![GitHub](https://img.shields.io/badge/GitHub-SC--Captioner-blue?logo=github)](https://github.com/zl2048/SC-Captioner)

All files are processed into llamafactory data format.

In `train_coco6k.json`, the "rejected" line means original gpt4 captions in RefinedCaps. They are not used in our self-correction training, but can be used for other purposes.