Speech Separation on Libri2Mix

10개 결과 · ⬇ CSV · JSON

SI-SDRi

Rank	Model	SI-SDRi	SDRi	Number of parameters (M)	SDR	Extra Training Data	Paper	Code	Year
1	MossFormer2 (w speed perturb)	22.2	–	–	–		MossFormer2: Combining Transformer and RNN-Free Recurrent Network for Enhanced Time-Domain Monaural Speech Separation	modelscope/ClearerVoice-Studio · alibabasglab/MossFormer2	2023
2	TF-Locoformer (M)	22.1	22.2	15	–		TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement	merlresearch/tf-locoformer	2024
3	MossFormer2 (w/o DM)	21.7	–	–	–		MossFormer2: Combining Transformer and RNN-Free Recurrent Network for Enhanced Time-Domain Monaural Speech Separation	modelscope/ClearerVoice-Studio · alibabasglab/MossFormer2	2023
4	Separate And Diffuse	21.5	–	–	–		Separate And Diffuse: Using a Pretrained Diffusion Model for Improving Source Separation		2023
5	WHYV	17.5	–	–	17.2458		Wanna hear your voice? A sample is all we need!		2024
6	TDANet Large	17.4	–	–	–		An efficient encoder-decoder architecture with top-down attention for speech separation	JusperLee/TDANet	2022
7	TDANet	16.9	–	–	–		An efficient encoder-decoder architecture with top-down attention for speech separation	JusperLee/TDANet	2022
8	Conv-Tasnet (Libri1Mix speech enhancement pre-trained)	14.1	14.6	–	–	✓	Stabilizing Label Assignment for Speech Separation by Self-supervised Pre-training	SungFeng-Huang/SSL-pretraining-separation	2020
9	Conv-Tasnet (Libri1Mix speech enhancement multi-task)	13.7	14.1	–	–	✓	Stabilizing Label Assignment for Speech Separation by Self-supervised Pre-training	SungFeng-Huang/SSL-pretraining-separation	2020
10	Conv-Tasnet	13.2	13.6	–	–		Stabilizing Label Assignment for Speech Separation by Self-supervised Pre-training	SungFeng-Huang/SSL-pretraining-separation	2020