HyperAIHyperAI
HomeConsoleDocsNewsPapersTutorialsDatasetsWikiSOTALLM ModelsGPU LeaderboardEvents
Search
About
HyperAIHyperAI

Command Palette

Search for a command to run...

Music Source Separation On Musdb18

Metrics

SDR (avg)
SDR (bass)
SDR (drums)
SDR (other)
SDR (vocals)

Results

Performance results of various models on this benchmark

Paper TitleRepository
Sparse HT Demucs (fine tuned)9.2010.4710.836.419.37Hybrid Transformers for Music Source Separation
Hybrid Transformer Demucs (f.t.)9.009.7810.086.429.20Hybrid Transformers for Music Source Separation
Band-Split RNN (semi-sup.)8.978.1610.157.0810.47Music Source Separation with Band-split RNN
TFC-TDF-UNet (v3)8.348.458.446.869.59Sound Demixing Challenge 2023 Music Demixing Track Technical Report: TFC-TDF-UNet v3
Band-Split RNN8.237.518.586.6210.21Music Source Separation with Band-split RNN
Hybrid Demucs7.728.678.585.598.04Hybrid Spectrogram and Waveform Source Separation
KUIELab-MDX-Net7.547.867.335.959.00KUIELab-MDX-Net: A Two-Stream Neural Network for Music Demixing
CDE-HTCN6.897.927.334.927.37Hierarchic Temporal Convolutional Network With Cross-Domain Encoder for Music Source Separation-
Attentive-MultiResUNet6.815.887.635.148.57An Efficient Short-Time Discrete Cosine Transform and Attentive MultiResUNet Framework for Music Source Separation-
DEMUCS (extra)6.797.607.584.697.29Music Source Separation in the Waveform Domain
CWS-PResUNet6.775.936.385.848.92CWS-PResUNet: Music Source Separation with Channel-wise Subband Phase-aware ResUNet
D3Net6.686.207.365.377.80D3Net: Densely connected multidilated DenseNet for music source separation
DLMNet6.477.297.054.626.91Contrastive Learning based Deep Latent Masking for Music Source Separation-
Conv-TasNet (extra)6.327.007.11-6.74Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation
UMXL6.3166.0157.1484.8897.213Open-Unmix - A Reference Implementation for Music Source Separation-
DEMUCS6.287.016.864.426.84Music Source Separation in the Waveform Domain
TAK26.045.406.814.807.16MMDenseLSTM: An efficient combination of convolutional and recurrent neural networks for audio source separation
D3Net6.015.257.014.537.24D3Net: Densely connected multidilated DenseNet for music source separation
Spleeter (MWF)5.915.516.714.026.86Spleeter: A Fast And State-of-the Art Music Source Separation Tool With Pre-trained Models-
LaSAFT+GPoCM5.885.635.684.877.33LaSAFT: Latent Source Attentive Frequency Transformation for Conditioned Source Separation
0 of 27 row(s) selected.
Music Source Separation On Musdb18 | SOTA | HyperAI