facebookresearch/AudioDec

Issues

• feature matching loss and adversarial loss rise steadily?

  #32 opened 3 months ago by Chengbin-Liang
  6

• [tensorboard] Show?

  #33 opened 3 months ago by a897456
  5

• [metrics] ViSQOL

  #34 opened 3 months ago by a897456
  3

• HiFiGAN is not supported for the analyzer?

  #31 opened 3 months ago by a897456
  6

• Is it missing some activation functions between some layers?

  #9 opened 4 months ago by BridgetteSong
  8

• The test results are different from those in the paper

  #12 opened 4 months ago by WzyCrush
  7

• VCTK dataset

  #16 opened 4 months ago by huazhi1024
  11

• vq_loss increase, not converge

  #17 opened 4 months ago by lixinghe1999
  7

• Table 4 and 5

  #25 opened 4 months ago by a897456
  2

• downsample_scales？

  #26 opened 4 months ago by a897456
  3

• Processing the same signal twice yields different results

  #27 opened 4 months ago by ilic-mezza
  3

• Speaker ID for real-time VC

  #28 opened 4 months ago by karthik19967829
  4

• How to obtain the encoded file

  #29 opened 4 months ago by XiangSheng0607
  1

• LOSS?

  #19 opened 6 months ago by a897456
  8

• Trying to train a low bitrate version

  #24 opened 5 months ago by McFlyy21
  2

• Fix the encoder

  #23 opened 6 months ago by a897456
  4

• projector(e)？

  #18 opened 6 months ago by a897456
  7

• quantizer dropout

  #20 opened 6 months ago by a897456
  3

• Baseline?

  #22 opened 6 months ago by a897456
  4

• LibriTTS dataset

  #21 opened 6 months ago by a897456
  2

• Perplexity blows up at start of training

  #8 opened 10 months ago by archie1993
  9

• Permissive license + patents

  #15 opened 8 months ago by kwinz
  0

• Some questions about CausalConvTranspose1d in conv_layer.py

  #11 opened 9 months ago by Jacksonroad
  3

• The trained models seem overfitted to their training sets?

  #7 opened 10 months ago by jpc
  3

• What data was used by the pretrained model of LibriTTS？

  #10 opened a year ago by jiaweiru
  1

• Increasing batch length

  #6 opened a year ago by anton-jeran
  2

• Samplerates' affect on hyperparameters

  #4 opened a year ago by fmac2000
  4

• unable to access exp link

  #5 opened a year ago by abhis-ek
  1

• Multi-GPU training

  #3 opened a year ago by jiaweiru
  2

• download link. needs. more. juice.

  #2 opened a year ago by ryanrussell
  1

• Relation / comparison to encodec

  #1 opened a year ago by vadimkantorov
  5