Home/Datasets/Freesound Dataset FSD50K - General Sound Dataset

Freesound Dataset FSD50K - General Sound Dataset

51,197 labeled audio samples covering 200 sound classes from Freesound.

About Dataset

This Zenodo record represents a research dataset archived and published through Zenodo, an open-access data repository operated by CERN. The dataset is accompanied by rich metadata, including author information, licensing, and version history, ensuring traceability and reproducibility. Zenodo records often host datasets related to academic publications, benchmarks, or supplementary experimental data, and are preserved for long-term access using persistent identifiers.

Purpose of Dataset

Datasets Hosted On Zenodo Are Used To Support Open Science, Reproducible Research, And Transparent Sharing Of Experimental Data. In Ai And Machine Learning Research, Such Datasets Are Commonly Used For Benchmarking, Validation, And Replication Of Published Results. They Also Serve As Reference Datasets For Academic Studies And Educational Purposes, Enabling Consistent Evaluation Across Research Groups.