TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/Conv-TasNet

Conv-TasNet

Reported on 19 benchmarks across 4 tasks · 1 paper · 13 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Audio6 results

  • Speech EnhancementonEARS-WHAM
    DNSMOS· 2018-09-20
    3.47
    best: 3.88 (SGMSE+)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech EnhancementonEARS-WHAM
    ESTOI· 2018-09-20
    0.7
    best: 0.73 (Schrödinger Bridge (PESQ loss))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech EnhancementonEARS-WHAM
    PESQ-WB· 2018-09-20
    2.31
    best: 3.09 (Schrödinger Bridge (PESQ loss))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech EnhancementonEARS-WHAM
    POLQA· 2018-09-20
    2.73
    best: 3.71 (Schrödinger Bridge (PESQ loss))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech EnhancementonEARS-WHAM
    SI-SDR· 2018-09-20
    16.93
    best: 17.85 (Schrödinger Bridge)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech EnhancementonEARS-WHAM
    SIGMOS· 2018-09-20
    2.69
    best: 3.44 (Schrödinger Bridge)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454

Music5 results

  • Music Source SeparationonMUSDB18
    SDR (other)· 2018-09-20
    4.37
    best: 7.08 (Band-Split RNN (semi-sup.))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Music Source SeparationonMUSDB18
    SDR (vocals)· 2018-09-20
    6.81
    best: 10.47 (Band-Split RNN (semi-sup.))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Music Source SeparationonMUSDB18
    SDR (avg)· 2018-09-20
    5.73
    best: 9.2 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Music Source SeparationonMUSDB18
    SDR (bass)· 2018-09-20
    5.66
    best: 10.47 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Music Source SeparationonMUSDB18
    SDR (drums)· 2018-09-20
    6.08
    best: 10.83 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454

Methodology5 results

  • 2D ClassificationonMUSDB18
    SDR (other)· 2018-09-20
    4.37
    best: 7.08 (Band-Split RNN (semi-sup.))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • 2D ClassificationonMUSDB18
    SDR (vocals)· 2018-09-20
    6.81
    best: 10.47 (Band-Split RNN (semi-sup.))
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • 2D ClassificationonMUSDB18
    SDR (avg)· 2018-09-20
    5.73
    best: 9.2 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • 2D ClassificationonMUSDB18
    SDR (bass)· 2018-09-20
    5.66
    best: 10.47 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • 2D ClassificationonMUSDB18
    SDR (drums)· 2018-09-20
    6.08
    best: 10.83 (Sparse HT Demucs (fine tuned))
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454

Speech3 results

  • Speech SeparationonWSJ0-2mix
    Number of parameters (M)· 2018-09-20
    5.1
    best: 59.4 (SepReformer-L)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech SeparationonWSJ0-2mix
    SDRi· 2018-09-20
    15.6
    best: 25.2 (TF-Locoformer (L) + DM)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454
  • Speech SeparationonWSJ0-2mix
    SI-SDRi· 2018-09-20
    15.3
    best: 25.1 (TF-Locoformer (L) + DM)
    SOTA
    Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech SeparationarXiv:1809.07454