ESC-50

The ESC-50 dataset is a benchmark that contains 2,000 environmental audio recordings across 50 categories, used to evaluate audio classification models.

Papers using ESC-50 (3)

Self-Supervised MultiModal Versatile Networks2020 · 195 citesONE-PEACE: Exploring One General Representation Model Toward Unlimited Modalities2023 · 42 citesEnhancing Audio-Language Models through Self-Supervised Post-Training with Text-Audio Pairs2024