Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
This project is mirrored from
https://github.com/lucidrains/audiolm-pytorch
. Pull mirroring updated
Sep 19, 2024
.
1.7.8
51a9ef35
·
minimize gradient syncs for soundstream distributed training
·
Nov 15, 2023
1.7.7
d6988ab2
·
1.7.7
·
Nov 15, 2023
1.7.6
044d5223
·
1.7.6
·
Nov 11, 2023
1.7.5
6e74cbad
·
when reconstructing from coarse transformer wrapper, handle variable lengthed...
·
Nov 11, 2023
1.7.4
591671a4
·
1.7.4
·
Nov 04, 2023
1.7.3
dc20c0e0
·
codebook size derived differently for residual FSQ
·
Nov 03, 2023
1.7.2
644f9919
·
add residual finite scalar quantizer
·
Nov 03, 2023
1.7.1
3669ec8b
·
default residual lfq to false, on reports it is performing poorly, residual fsq to be added
·
Nov 03, 2023
1.7.0
a8d9e873
·
get ready to test residual LFQ for soundstream, setup some convenience code for wandb experiments
·
Nov 03, 2023
1.6.7
cdeedb9d
·
helper method and accessor off soundstream trainer
·
Oct 30, 2023
1.6.6
94cbbb23
·
patch
·
Oct 30, 2023
1.6.5
0124c4bc
·
sanity check, clearly demonstrate tokenization of audio ability using residual LFQ
·
Oct 28, 2023
1.6.4
c4f2ef56
·
address
https://github.com/lucidrains/audiolm-pytorch/issues/242
·
Oct 27, 2023
1.6.3
1b4d80f9
·
allow for local attention to attend to nothing
·
Oct 24, 2023
1.6.2
8fd178d4
·
increase default codebook size
·
Oct 24, 2023
1.6.1
29a57ef7
·
allow for grouped residual LFQ
·
Oct 23, 2023
1.6.0
c748fcdb
·
integrate residual lookup free quantization
·
Oct 22, 2023
1.5.7
4bc50b02
·
address padding in acoustic tokens when sampling from fine transformer...
·
Oct 10, 2023
1.5.6
17ebcae1
·
address
https://github.com/lucidrains/audiolm-pytorch/issues/235
·
Oct 06, 2023
1.5.5
b394aecb
·
more barriers
·
Oct 05, 2023
Prev
1
2
3
4
5
6
…
17
Next