Speech Separation on WSJ0-2mix

Metric: Number of parameters (M) (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	Number of parameters (M)▼	Extra Data	Paper	Date↕	Code
1	SepReformer-L	59.4	No	Separate and Reconstruct: Asymmetric Encoder-Dec...	2024-06-10	Code
2	MossFormer2 (L)	55.7	No	-	-	Code
3	MossFormer (L) + DM	42.1	No	MossFormer: Pushing the Performance Limit of Mon...	2023-02-23	Code
4	SPGM + DM	26.2	No	SPGM: Prioritizing Local Features for enhanced s...	2023-09-22	Code
5	SPGM	26.2	No	SPGM: Prioritizing Local Features for enhanced s...	2023-09-22	Code
6	TF-Locoformer (L) + DM	22.5	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
7	TF-Locoformer (L)	22.5	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
8	TF-Locoformer (M) + DM	15	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
9	TF-Locoformer (M)	15	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
10	Conv-TasNet	5.1	No	Conv-TasNet: Surpassing Ideal Time-Frequency Mag...	2018-09-20	Code
11	TF-Locoformer (S) + DM	5	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
12	TF-Locoformer (S)	5	No	TF-Locoformer: Transformer with Local Modeling b...	2024-08-06	Code
13	Deformable TCN + Dynamic Mixing	3.6	No	Deformable Temporal Convolutional Networks for M...	2022-10-27	Code
14	Deformable TCN + Shared Weights + Dynamic Mixing	1.3	No	Deformable Temporal Convolutional Networks for M...	2022-10-27	Code

#1SepReformer-LSOTA
59.4
Number of parameters (M)· 2024-06-10
Separate and Reconstruct: Asymmetric Encoder-Decoder for Speech Separation Code
#2MossFormer2 (L)
55.7
Number of parameters (M)
No paperCode
#3MossFormer (L) + DMSOTA
42.1
Number of parameters (M)· 2023-02-23
MossFormer: Pushing the Performance Limit of Monaural Speech Separation using Gated Single-Head Transformer with Convolution-Augmented Joint Self-Attentions Code
#4SPGM + DM
26.2
Number of parameters (M)· 2023-09-22
SPGM: Prioritizing Local Features for enhanced speech separation performance Code
#5SPGM
26.2
Number of parameters (M)· 2023-09-22
SPGM: Prioritizing Local Features for enhanced speech separation performance Code
#6TF-Locoformer (L) + DM
22.5
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#7TF-Locoformer (L)
22.5
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#8TF-Locoformer (M) + DM
15
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#9TF-Locoformer (M)
15
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#10Conv-TasNetSOTA
5.1
Number of parameters (M)· 2018-09-20
Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation Code
#11TF-Locoformer (S) + DM
5
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#12TF-Locoformer (S)
5
Number of parameters (M)· 2024-08-06
TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement Code
#13Deformable TCN + Dynamic Mixing
3.6
Number of parameters (M)· 2022-10-27
Deformable Temporal Convolutional Networks for Monaural Noisy Reverberant Speech Separation Code
#14Deformable TCN + Shared Weights + Dynamic Mixing
1.3
Number of parameters (M)· 2022-10-27
Deformable Temporal Convolutional Networks for Monaural Noisy Reverberant Speech Separation Code