Software Open Access

Cosentino/ConvTasNet_LibriMix_sep_noisy

Joris Cosentino

Description:

This model was trained by Joris Cosentino using the librimix recipe in Asteroid. It was trained on the sep_noisy task of the LibriMix dataset.

 

 

Training config:

  • data:
    • n_src: 2
    • sample_rate: 8000
    • segment: 3
    • task: sep_noisy
    • train_dir: data/wav8k/min/train-360
    • valid_dir: data/wav8k/min/dev
  • filterbank:
    • kernel_size: 16
    • n_filters: 512
    • stride: 8
  • masknet:
    • bn_chan: 128
    • hid_chan: 512
    • mask_act: relu
    • n_blocks: 8
    • n_repeats: 3
    • skip_chan: 128
  • optim:
    • lr: 0.001
    • optimizer: adam
    • weight_decay: 0.0
  • positional arguments:
  • training:
    • batch_size: 24
    • early_stop: True
    • epochs: 200
    • half_lr: True
    • num_workers: 4

 

Results:

  • si_sdr: 9.944424856077259
  • si_sdr_imp: 11.939395359731192
  • sdr: 10.701526190782072
  • sdr_imp: 12.481757547845662
  • sir: 22.633644975545575
  • sir_imp: 22.45666740833025
  • sar: 11.131644100944868
  • sar_imp: 4.248489589311784
  • stoi: 0.852048619949357
  • stoi_imp: 0.2071994899565506

 

License notice:

This work "ConvTasNet_Librisep_noisyMix_sepnoisy" is a derivative of LibriSpeech ASR corpus by Vassil Panayotov, used under CC BY 4.0; of The WSJ0 Hipster Ambient Mixtures dataset by Whisper.ai, used under CC BY-NC 4.0 (Research only). "ConvTasNet_Librisep_noisyMix_sepnoisy" is licensed under Attribution-ShareAlike 3.0 Unported by Joris Cosentino.

Files (20.3 MB)
Name Size
model.pth
md5:817816fc622dabbe9ce51d431780725e
20.3 MB Download
479
493
views
downloads
All versions This version
Views 479479
Downloads 493493
Data volume 10.0 GB10.0 GB
Unique views 430430
Unique downloads 362362

Share

Cite as