FSD50K

Freesound Database 50K

Dataset Information
Modalities
Audio
License
Other (Attribution)
Homepage
Official Website
Contents

Overview

Freesound Dataset 50k (or FSD50K for short) is an open dataset of human-labeled sound events containing 51,197 Freesound clips unequally distributed in 200 classes drawn from the AudioSet Ontology. FSD50K has been created at the Music Technology Group of Universitat Pompeu Fabra. It consists mainly of sound events produced by physical sound sources and production mechanisms, including human sounds, sounds of things, animals, natural sounds, musical instruments and more.

Source: https://zenodo.org/record/4060432
Image Source: https://labs.freesound.org/datasets/

Variants: FSD50K

Associated Benchmarks

This dataset is used in 2 benchmarks:

Recent Benchmark Submissions

Task Model Paper Date
Audio Classification MATPAC (SSL Model) Masked Latent Prediction and Classification … 2025-02-17
Audio Classification LHGNN LHGNN: Local-Higher Order Graph Neural … 2025-01-07
Audio Classification MN Dynamic Convolutional Neural Networks as … 2023-10-24
Audio Classification DyMN-L Dynamic Convolutional Neural Networks as … 2023-10-24
Audio Classification ONE-PEACE ONE-PEACE: Exploring One General Representation … 2023-05-18
Environmental Sound Classification [ABT] AudioNTT Audio Barlow Twins: Self-Supervised Audio … 2022-09-28
Audio Classification Temporal Knowledge Distillation for On-device Audio Classification Temporal Knowledge Distillation for On-device … 2021-10-27
Audio Classification PaSST-N-S Efficient Training of Audio Transformers … 2021-10-11
Audio Classification PaSST-S Efficient Training of Audio Transformers … 2021-10-11
Audio Classification Large 6-Layer Transformer with Pooling Audio Transformers 2021-05-01
Audio Classification PSLA PSLA: Improving Audio Tagging with … 2021-02-02

Research Papers

Recent papers with results on this dataset:

External Links: