---
license: cc-by-4.0
dataset_info:
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: transcription
    dtype: string
  - name: summary
    dtype: string
  - name: summary1
    dtype: string
  - name: summary2
    dtype: string
  - name: summary3
    dtype: string
  splits:
  - name: core
    num_bytes: 17683719490.0
    num_examples: 50000
  - name: duc2003
    num_bytes: 244384744.0
    num_examples: 624
  - name: validation
    num_bytes: 342668783.0
    num_examples: 1000
  - name: test
    num_bytes: 1411039659.0
    num_examples: 4000
  download_size: 19837902893
  dataset_size: 19681812676.0
configs:
- config_name: default
  data_files:
  - split: core
    path: data/core-*
  - split: duc2003
    path: data/duc2003-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---
# Mega-SSum
- A large-scale English *sentence-wise speech summarization* (Sen-SSum) dataset
  - Consists of 3.8M+ synthesized speech, transcription, summary triplets
  - Derived from the Gigaword dataset [Rush+2015](https://aclanthology.org/D15-1044/)

# Overview
- The dataset is divided into five splits: train/core/dev/eval/duc2003. (See below table)
  - We added a new evaluation split "*test*" for in-domain evaluation.
  - The train split is here: [MegaSSum(train)](https://huggingface.co/datasets/komats/mega-ssum-train).

| orig. data | split     | #samples  | #speakers | total dur. (hrs) | ave. dur. (sec) | CR* (%) |
|:----------:|:---------:|:---------:|:---------:|:----------------:|:---------------:|--------:|
| Gigaword   | train     | 3,800,000 | 2,559     | 11,678.2         | 11.1            | 26.2   |
| Gigaword   | core      | 50,000    | 2,559     | 154.6            | 11.1            | 25.8   |
| Gigaword   | valid     | 1,000     | 96        | 3.0              | 10.7            | 25.1   |
| Gigaword   | test      | 4,000     | 80        | 12.5             | 11.2            | 24.1   |
| DUC2003    | duc2003   | 624       | 80        | 2.1              | 12.2            | 27.5   |

*CR (compression rate, %) = #words in summary / #words in transcription * 100. Lower is shorter summary.

# Notes
- The core set is identical to the first 50k samples of the train split.
  - You may train your model and report the results only with the core set because the train split is very large.
  - Using the entire train split is generally not recommended unless there are special reasons (e.g., to investigate the upper bound).
- The duc2003 split has four reference summaries for each speech. You can report the best score from 4 scores.
- Spoken sentences were generated using VITS [Kim+2021](https://proceedings.mlr.press/v139/kim21f.html) trained with LibriTTS-R [Koizumi+2023](https://www.isca-archive.org/interspeech_2023/koizumi23_interspeech.html).
- More details and some experiments on this dataset can be found [here](https://www.isca-archive.org/interspeech_2024/matsuura24_interspeech.html#).

# Citation
- This dataset [Matsuura+2024](https://www.isca-archive.org/interspeech_2024/matsuura24_interspeech.html):
  ```
  @inproceedings{matsuura24_interspeech,
    title     = {{Sentence-wise Speech Summarization}: Task, Datasets, and End-to-End Modeling with LM Knowledge Distillation},
    author    = {Kohei Matsuura and Takanori Ashihara and Takafumi Moriya and Masato Mimura and Takatomo Kano and Atsunori Ogawa and Marc Delcroix},
    year      = {2024},
    booktitle = {Interspeech 2024},
    pages     = {1945--1949},
  }
  ```

- The Gigaword dataset [Rush+2015](https://aclanthology.org/D15-1044/):
  ```
  @article{Rush_2015,
     title={A Neural Attention Model for Abstractive Sentence Summarization},
     journal={Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},
     author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},
     year={2015}
  }
  ```

- VITS TTS [Kim+2021](https://proceedings.mlr.press/v139/kim21f.html):
  ```
  @InProceedings{pmlr-v139-kim21f,
    title = 	 {Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech},
    author =       {Kim, Jaehyeon and Kong, Jungil and Son, Juhee},
    booktitle = 	 {Proceedings of the 38th International Conference on Machine Learning},
    pages = 	 {5530--5540},
    year = 	 {2021},
  }
  ```

- LibriTTS-R [Koizumi+2023](https://www.isca-archive.org/interspeech_2023/koizumi23_interspeech.html):
  ```
  @inproceedings{koizumi23_interspeech,
    author={Yuma Koizumi and Heiga Zen and Shigeki Karita and Yifan Ding and Kohei Yatabe and Nobuyuki Morioka and Michiel Bacchiani and Yu Zhang and Wei Han and Ankur Bapna},
    title={{LibriTTS-R}: A Restored Multi-Speaker Text-to-Speech Corpus},
    year=2023,
    booktitle={Proc. INTERSPEECH 2023},
    pages={5496--5500},
  }
  ```