AI Paper Atlas

Research map / Benchmarks / ESC-50

ESC-50 leaderboard

36 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on ESC-50. Best accuracy: 99.1 (Ours, CVPR 2024).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

ESC-50 · accuracy

36 papers, each with its best reported number for its own method; higher is better.

#accuracy ↑MethodPaperTopic
199.1OursOmniVec2 - A Novel Transformer based Network for Large Scale Multimodal and Multitask Learning
CVPR 2024 · Table 4
Cross-modal learning
298.1BEATs_iter3+ (with AS-2M supervised training)
BEATs
BEATs: Audio Pre-Training with Acoustic Tokenizers
ICML 2023 · Table 2
Self-supervised and contrastive
397.5Full TF (Ours), AudioSet full set pretrainedDTF-AT: Decoupled Time-Frequency Audio Transformer for Event Classification
AAAI 2024 · Table 2
CNN design
496.8CMKD (AudioSet pretrain + KD)
AST-Base
CMKD: CNN/Transformer-Based Cross-Model Knowledge Distillation for Audio Classification
TPAMI 2025 · text
Knowledge distillation
596.2SSLAM
ViT-Base
SSLAM: Enhancing Self-Supervised Models with Audio Mixtures for Polyphonic Soundscapes
ICLR 2025 · Table: Evaluation on audio SSL
Self-supervised and contrastive
696EquiAV
ViT-B/16
EquiAV: Leveraging Equivariance for Audio-Visual Contrastive Learning
ICML 2024 · Audio classification table
Cross-modal learning
794.6AudioMAE + R-BN
ViT-B (AudioSet-2M pretraining)
Structured-Noise Masked Modeling for Video, Audio and Beyond
ECCV 2026 · Table 4
Video segmentation and tracking
894.4MAViL
ViT-B
MAViL: Masked Audio-Video Learners
NeurIPS 2023 · Table (audio-only tasks)
Video segmentation and tracking
994.1Audio-MAE (local)
ViT-B
Masked Autoencoders that Listen
NeurIPS 2022 · Table 2
Self-supervised and contrastive
1093.6InternVL_IB++ (AT E.)
ImageBind + InternVL
FreeBind: Free Lunch in Unified Multimodal Space via Knowledge Fusion
ICML 2024 · Results of zero-shot classific
Cross-modal learning
26 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.