ESC-50

Overview edit

Variants: ESC-50

Associated Benchmarks

This dataset is used in 3 benchmarks:

• Audio Classification - Metrics: Top-1 Accuracy, PRE-TRAINING DATASET, Accuracy (5-fold)
• Environmental Sound Classification - Metrics: Accuracy
• Image Classification - Metrics: Top 1 Accuracy

Recent Benchmark Submissions

Research Papers

Recent papers with results on this dataset:

• M2D2: Exploring General-purpose Audio-Language Representations Beyond CLAP (2025) -
• Masked Latent Prediction and Classification for Self-Supervised Audio Representation Learning (2025) -
• LHGNN: Local-Higher Order Graph Neural Networks For Audio Classification and Tagging (2025) -
• Performance of Gaussian Mixture Model Classifiers on Embedded Feature Spaces (2024) -
• M2D-CLAP: Masked Modeling Duo Meets CLAP for Learning General-purpose Audio-Language Representation (2024) -
• Masked Modeling Duo: Towards a Universal Audio Pre-training Framework (2024) -
• InternVideo2: Scaling Foundation Models for Multimodal Video Understanding (2024) -
• EAT: Self-Supervised Pre-Training with Efficient Audio Transformer (2024) -
• OmniVec: Learning robust representations with cross modal sharing (2023) -
• Dynamic Convolutional Neural Networks as Efficient Pre-trained Audio Models (2023) -
• BEATs: Audio Pre-Training with Acoustic Tokenizers (2022) -
• Efficient Large-scale Audio Tagging via Transformer-to-CNN Knowledge Distillation (2022) -
• Learning Rate Curriculum (2022) -
• End-to-End Audio Strikes Back: Boosting Augmentations Towards An Efficient Audio Classification Network (2022) -
• SepTr: Separable Transformer for Audio Spectrogram Processing (2022) -
• HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection (2022) -
• AudioCLIP: Extending CLIP to Image, Text and Audio (2021) -
• AST: Audio Spectrogram Transformer (2021) -
• Multi-Format Contrastive Learning of Audio Representations (2021) -
• Environmental Sound Classification on the Edge: A Pipeline for Deep Acoustic Networks on Extremely Resource-Constrained Devices (2021) -