HEAR Leaderboard

Model	URL	Submission Date	Beehive	Beijing Opera	CREMA-D	DCASE 2016	ESC-50	FSD50K	GTZAN Genre	GTZAN Music/Speech	Gunshot	Libricount	Maestro 5h	Mridangam Stroke	Mridangam Tonic	NSynth Pitch 50h	NSynth Pitch 5h	Speech commands 5h	Speech commands full	Vocal Imitation	VoxLingua107 top 10

Model	Submission Date	Beehive	Beijing Opera	CREMA-D	DCASE 2016	ESC-50	FSD50K	GTZAN Genre	GTZAN Music/Speech	Gunshot	Libricount	Maestro 5h	Mridangam Stroke	Mridangam Tonic	NSynth Pitch 50h	NSynth Pitch 5h	Speech commands 5h	Speech commands full	Vocal Imitation	VoxLingua107 top 10
RedRice ced_base	2023-09-26	0.4835	0.9660	0.6910	0.9219	0.9665	0.6548	0.8860	0.9436	0.8929	0.6785	0.1476	0.9743	0.9655	0.8281	0.6820	0.8693	0.8967	0.2269	0.3857
GURA Fuse Cat H+w+C	HEAR 2021		0.9660	0.7474	0.8260	0.7335	0.4197	0.8050	0.9282	0.9345	0.6972	0.4413	0.9725	0.9235	0.8852	0.8460	0.9605	0.9677	0.1969	0.7202
RedRice ced_small	2023-09-26	0.5170	0.9660	0.6664	0.9163	0.9595	0.6433	0.8950	0.9122	0.9345	0.6559	0.1096	0.9682	0.9394	0.7995	0.6020	0.8092	0.8519	0.2192	0.3653
GURA Fuse Cat H+w+C (time)	HEAR 2021		0.9618	0.7427	0.8260	0.6535	0.3742	0.7600	0.9436	0.9048	0.6591	0.4413	0.9753	0.9243	0.8909	0.8540	0.9509	0.9681	0.2152	0.6286
RedRice ced_mini	2023-09-26	0.5917	0.9618	0.6526	0.9066	0.9535	0.6388	0.9030	0.9449	0.8601	0.6402	0.08289	0.9656	0.9332	0.7520	0.5560	0.7738	0.8196	0.2037	0.3467
Logitech AI SERAB BYOL-S	HEAR 2021	0.5487	0.9533	0.6569	0.6421	0.8050	0.5088	0.8370	0.9385	0.8571	0.7853	0.007860	0.9726	0.9285	0.7116	0.3960	0.9137	0.9481	0.1598	0.4578
CP-JKU mn40_as (e1+l+all_se)	2023-01-05	0.5027	0.9534	0.6400	0.8130	0.9615	0.6312	0.9080	0.9840	0.9107	0.7253	0.01862	0.9713	0.9047	0.6355	0.3040	0.7767	0.8472	0.2019	0.3179
GURA Fuse Hubert	HEAR 2021		0.9493	0.7521	0.8259	0.7435	0.4132	0.7960	0.9359	0.9286	0.6834	0.1657	0.9738	0.9090	0.6882	0.3820	0.9468	0.9571	0.1848	0.7140
GURA Cat H+w+C	HEAR 2021		0.9363	0.6394	0.6808	0.5110	0.3144	0.7220	0.9609	0.8810	0.6390	0.4691	0.9378	0.8591	0.8968	0.8660	0.9270	0.9429	0.1114	0.4599
OpenL3	HEAR 2021	0.6040	0.9746	0.5497	0.8328	0.7505	0.4470	0.8790	0.9686	0.9494	0.6414	0.01650	0.9666	0.9369	0.7310	0.5600	0.6796	0.7634	0.07812	0.3313
GURA Fuse wav2vec2	HEAR 2021		0.9449	0.6924	0.7983	0.6950	0.4028	0.7930	0.9532	0.9673	0.6526	0.1110	0.9623	0.8376	0.6059	0.3300	0.9571	0.9687	0.1740	0.7058
GURA Avg H+w+C	HEAR 2021		0.9448	0.5473	0.6236	0.4495	0.2636	0.7060	0.9365	0.8571	0.6309	0.4599	0.9144	0.8373	0.8963	0.8620	0.8225	0.8808	0.08415	0.3210
CP-JKU PaSST 2lvl+mel	HEAR 2021		0.9660	0.6104	0.9254	0.9475	0.6409	0.8830	0.9769	0.9405	0.6601		0.9650	0.8194	0.5409	0.2560	0.6810	0.6387	0.1820	0.2593
CP-JKU PaSST 2lvl	HEAR 2021		0.9660	0.6104	0.9132	0.9475	0.6409	0.8830	0.9769	0.9405	0.6601		0.9650	0.8194	0.5409	0.2560	0.6810	0.6387	0.1820	0.2593
GURA Avg Hubert+Crepe	HEAR 2021		0.9323	0.5397	0.6104	0.4365	0.2528	0.6980	0.9455	0.8452	0.6224	0.4624	0.9170	0.8309	0.8966	0.8780	0.7372	0.8233	0.07895	0.2932
RedRice EfficientNet-B2	HEAR 2021	0.5332	0.9535	0.5746	0.7901	0.9345	0.6071	0.8780	0.9679	0.8780	0.6509	0.0001910	0.9485	0.8432	0.3914	0.1680	0.5734	0.6757	0.1385	0.2551
CP-JKU PaSST base	HEAR 2021		0.9660	0.6104	0.7879	0.9475	0.6409	0.8830	0.9769	0.9405	0.6601		0.9650	0.8194	0.5409	0.2560	0.6810	0.6387	0.1820	0.2593
GURA Cat wav2vec2+crepe	HEAR 2021		0.9198	0.4598	0.5855	0.3430	0.2341	0.6810	0.9378	0.8333	0.5694	0.4626	0.8977	0.8226	0.8990	0.8680	0.8853	0.9188	0.07640	0.3097
wav2vec2	HEAR 2021		0.9067	0.6562	0.6630	0.5610	0.3417	0.7800	0.9462	0.8482	0.6921	0.03290	0.9432	0.8283	0.6530	0.4020	0.8382	0.8785	0.08006	0.4928
GURA Hubert	HEAR 2021		0.9447	0.6897	0.5841	0.6035	0.3136	0.7350	0.9128	0.9315	0.6458	0.006769	0.9528	0.8499	0.4288	0.1840	0.9530	0.9540	0.1540	0.6368
GURA Cat Hubert+wav2vec2	HEAR 2021		0.9363	0.6982	0.4524	0.5860	0.3225	0.7340	0.9359	0.8690	0.6273	0.003479	0.9511	0.8018	0.4276	0.1980	0.9360	0.9286	0.1659	0.7058
IUT-CSE MLP (keyword)	HEAR 2021	0.7597	0.9109	0.4243	0.5179	0.3675	0.1868	0.5540	0.8891	0.9315	0.4505	0.06481	0.9686	0.9415	0.6047	0.4400	0.9763	0.9779	0.05608	0.1811
GURA Avg Hubert+wav2vec2	HEAR 2021		0.9405	0.6992	0.4170	0.5865	0.3177	0.7460	0.9282	0.8095	0.6227	0.004417	0.9464	0.7993	0.4148	0.1980	0.9074	0.9519	0.1649	0.6904
Sony UDONS ViT	HEAR 2021	0.8780	0.9279	0.4406	0.6682	0.4010		0.6810	0.8994	0.8661	0.4879	0.2386	0.8796	0.7301	0.7946	0.6920	0.4789	0.5313	0.06830	0.2243
CP-JKU mn04_as (e1+l)	2023-01-05	0.4349	0.9279	0.4913	0.6499	0.9180	0.5668	0.8620	0.9923	0.8750	0.5502	0.003697	0.9530	0.7839	0.2818	0.1260	0.4818	0.4577	0.1439	0.2386
Stellenbosch LSL DBERT	HEAR 2021	0.6972	0.9191	0.5215	0.2462	0.5315	0.2635	0.6740	0.9686	0.8095	0.5843	0.0001635	0.8352	0.6848	0.7374	0.5240	0.5656	0.6305	0.1611	0.2459
CREPE	HEAR 2021	0.5932	0.9281	0.3828	0.5037	0.3005	0.1593	0.6450	0.9288	0.8631	0.4986	0.4011	0.8977	0.8236	0.9000	0.8700	0.1800	0.2112	0.05133	0.1420
CVSSP PANNS	HEAR 2021	0.4457	0.9112	0.5550	0.000	0.9085		0.8600	0.9923	0.7976	0.6516	0.000	0.9391	0.8244	0.3012	0.1480	0.5601	0.6182	0.1270	0.2438
Descript/MARL Wav2CLIP	HEAR 2021	0.7698	0.9363	0.5116	0.000	0.7590	0.3617	0.7480	0.9455	0.9286	0.5276	0.000	0.9471	0.8289	0.4389	0.2300	0.3160	0.3466	0.08334	0.1924
Kuroyanagi hearline	HEAR 2021		0.9281	0.4800		0.4245	0.1781	0.6540	0.9308	0.5179	0.4979	0.001639	0.9094	0.7834	0.5894	0.3940	0.4783	0.5798	0.03982	0.1801
AMAAI wav2vec2 music+speech	HEAR 2021		0.8261	0.3914	0.5098	0.5110	0.1980	0.6050	0.9071	0.8571	0.4318	0.002607	0.7595	0.4149	0.1759	0.1820	0.7436	0.5231	0.06415	0.1686
IUT-CSE MLP (audio)	HEAR 2021	0.5353	0.7284	0.4199	0.05225	0.1605	0.08624	0.4080	0.7737	0.7857	0.3664	0.02368	0.8932	0.7781	0.6076	0.3860	0.3920	0.5348	0.03759	0.1286
Soundsensing YAMNet	HEAR 2021	0.4665	0.9405	0.4533	0.008316	0.8375		0.8470	0.9686	0.7321	0.6526	0.000			0.3209	0.1580	0.2888	0.4104	0.08471	0.2016
AMAAI Lab wav2vec2+DDSP	HEAR 2021		0.6349		0.5070	0.2795	0.2678			0.5833	0.5103	0.0003349	0.6531	0.3572	0.3030	0.1720	0.4871	0.4904	0.07238	0.1060

Showing 1 to 34 of 34 entries

Leaderboard Heatmap

Primary score of leaderboard models on each HEAR task. Normalized scores are used to show the heat-value of each cell.

[EPS]

Leaderboard Heatmap

Task Correlations

Task versus task correlation scores, based upon normalized scores. Only the highest and lowest correlations are displayed. Cells are sorted to minimize the traveling salesperson distance, mapping correlations [-1, +1] to distances [+2, 0].

[EPS]

Task Correlations

Model Correlations

Model versus model correlation scores, based upon normalized scores. Only the highest and lowest correlations are displayed. Cells are sorted to minimize the traveling salesperson distance, mapping correlations [-1, +1] to distances [+2, 0].

[EPS]

Model Correlations