---
license: mit
language:
- en
pretty_name: anon
---
# aha-annotationsv1
## Dataset Description

This repo contains the dataset **anon-annotationsv1**, which is used for training **anon**, and benchmarks for evaluating **anon**. The data distribution of anon-annotationsv1 is as follows:


<!-- - HIHD
  - [HIHD](https://github.com/MRHiSum/MR.HiSum/tree/main): 31892 examples (not all of them used)
- Dense Captioning
  - [Shot2Story](https://github.com/bytedance/Shot2Story): 36949 examples from human_anno subset
  - [COIN](https://coin-dataset.github.io/): 4574 examples from the train set with 2-4 minutes videos
- Multi-Answer Grounded Video Question Answering (MAGQA)
  - The proposed dataset for Multi-Answer Grounded Video Question Answering (MAGQA), **Shot2Story-MAGQA-39k**, is also included in this repository. Its training set is `shot2story/annotations/magqa_train-0.25_0.5-earlier.json`, and its test set is `shot2story/annotations/magqa_test.json`. This dataset is generated from the [MMDuet](https://huggingface.co/datasets/wangyueqian/MMDuetIT) work, please refer to their work for the details. -->

Please refer our github page for the usage.

## Related Resources