---
configs:
- config_name: audioset500k
  data_files:
  - split: train
    path: audioset500k.json
- config_name: balanced_train
  data_files:
  - split: train
    path: balanced_train.json
- config_name: eval
  data_files:
  - split: test
    path: eval.json
- config_name: unbalanced_train_part00
  data_files: unbalanced_train_part00.json
  # dataset_size: 46940
- config_name: unbalanced_train_part01
  data_files: unbalanced_train_part01.json
  # dataset_size: 47052
- config_name: unbalanced_train_part02
  data_files: unbalanced_train_part02.json
  # dataset_size: 46923
- config_name: unbalanced_train_part03
  data_files: unbalanced_train_part03.json
  # dataset_size: 46952
- config_name: unbalanced_train_part04
  data_files: unbalanced_train_part04.json
  # dataset_size: 46916
- config_name: unbalanced_train_part05
  data_files: unbalanced_train_part05.json
  # dataset_size: 47011
- config_name: unbalanced_train_part06
  data_files: unbalanced_train_part06.json
  # dataset_size: 46964
- config_name: unbalanced_train_part07
  data_files: unbalanced_train_part07.json
  # dataset_size: 46915
- config_name: unbalanced_train_part08
  data_files: unbalanced_train_part08.json
  # dataset_size: 46927
- config_name: unbalanced_train_part09
  data_files: unbalanced_train_part09.json
  # dataset_size: 46839
- config_name: unbalanced_train_part10
  data_files: unbalanced_train_part10.json
  # dataset_size: 46862
- config_name: unbalanced_train_part11
  data_files: unbalanced_train_part11.json
  # dataset_size: 46836
- config_name: unbalanced_train_part12
  data_files: unbalanced_train_part12.json
  # dataset_size: 46865
- config_name: unbalanced_train_part13
  data_files: unbalanced_train_part13.json
  # dataset_size: 46800
- config_name: unbalanced_train_part14
  data_files: unbalanced_train_part14.json
  # dataset_size: 46837
- config_name: unbalanced_train_part15
  data_files: unbalanced_train_part15.json
  # dataset_size: 46824
- config_name: unbalanced_train_part16
  data_files: unbalanced_train_part16.json
  # dataset_size: 46813
- config_name: unbalanced_train_part17
  data_files: unbalanced_train_part17.json
  # dataset_size: 46771
- config_name: unbalanced_train_part18
  data_files: unbalanced_train_part18.json
  # dataset_size: 46875
- config_name: unbalanced_train_part19
  data_files: unbalanced_train_part19.json
  # dataset_size: 46885
- config_name: unbalanced_train_part20
  data_files: unbalanced_train_part20.json
  # dataset_size: 46884
- config_name: unbalanced_train_part21
  data_files: unbalanced_train_part21.json
  # dataset_size: 46736
- config_name: unbalanced_train_part22
  data_files: unbalanced_train_part22.json
  # dataset_size: 46832
- config_name: unbalanced_train_part23
  data_files: unbalanced_train_part23.json
  # dataset_size: 46823
- config_name: unbalanced_train_part24
  data_files: unbalanced_train_part24.json
  # dataset_size: 46795
- config_name: unbalanced_train_part25
  data_files: unbalanced_train_part25.json
  # dataset_size: 46740
- config_name: unbalanced_train_part26
  data_files: unbalanced_train_part26.json
  # dataset_size: 46765
- config_name: unbalanced_train_part27
  data_files: unbalanced_train_part27.json
  # dataset_size: 46708
- config_name: unbalanced_train_part28
  data_files: unbalanced_train_part28.json
  # dataset_size: 46736
- config_name: unbalanced_train_part29
  data_files: unbalanced_train_part29.json
  # dataset_size: 46819
- config_name: unbalanced_train_part30
  data_files: unbalanced_train_part30.json
  # dataset_size: 46694
- config_name: unbalanced_train_part31
  data_files: unbalanced_train_part31.json
  # dataset_size: 46735
- config_name: unbalanced_train_part32
  data_files: unbalanced_train_part32.json
  # dataset_size: 46731
- config_name: unbalanced_train_part33
  data_files: unbalanced_train_part33.json
  # dataset_size: 46627
- config_name: unbalanced_train_part34
  data_files: unbalanced_train_part34.json
  # dataset_size: 46740
- config_name: unbalanced_train_part35
  data_files: unbalanced_train_part35.json
  # dataset_size: 46866
- config_name: unbalanced_train_part36
  data_files: unbalanced_train_part36.json
  # dataset_size: 46758
- config_name: unbalanced_train_part37
  data_files: unbalanced_train_part37.json
  # dataset_size: 46751
- config_name: unbalanced_train_part38
  data_files: unbalanced_train_part38.json
  # dataset_size: 46750
- config_name: unbalanced_train_part39
  data_files: unbalanced_train_part39.json
  # dataset_size: 46700
- config_name: unbalanced_train_part40
  data_files: unbalanced_train_part40.json
  # dataset_size: 39137
task_categories:
- audio-classification
tags:
- audioset
size_categories:
- 100M<n<1B
---

# AudioSet

AudioSet<sup>[1]</sup> consists of an expanding ontology of 527 audio event classes and a collection of 2M human-labelled 10-second sound clips drawn from YouTube. 
Some clips are missing on YouTube, so the number of files downloaded is different from time to time. 
This repository contains 20550 / 22160 of the balanced train set, 1913637 / 2041789 of the unbalanced train set (separated into 41 parts), and 18887 / 20371 of the evaluation set.
The pre-process script can be found at qiuqiangkong's [github](https://github.com/qiuqiangkong/audioset_tagging_cnn)<sup>[2]</sup>.

To improve training efficiency, we add a slightly more balanced subset AudioSet500K<sup>[3]</sup>.

## References

1. Gemmeke, Jort F., et al., Audio set: An ontology and human-labeled dataset for audio events, 2017
2. Kong, Qiuqiang, et al., Panns: Large-scale pretrained audio neural networks for audio pattern recognition, 2020
3. Nagrani, Arsha, et al., Attention bottlenecks for multimodal fusion, 2021