Releases: lucidrains/audiolm-pytorch
Releases · lucidrains/audiolm-pytorch
1.7.6
1.7.6
1.7.5
when reconstructing from coarse transformer wrapper, handle variable …
1.7.4
1.7.4
1.7.3
codebook size derived differently for residual FSQ
1.7.2
add residual finite scalar quantizer
1.7.1
default residual lfq to false, on reports it is performing poorly, re…
1.7.0
get ready to test residual LFQ for soundstream, setup some convenienc…
1.6.7
helper method and accessor off soundstream trainer
1.6.6
patch
1.6.5
sanity check, clearly demonstrate tokenization of audio ability using…