0.10.14

0.10.14

0.10.12

address https://github.com/lucidrains/vector-quantize-pytorch/issues/33

0.10.11

prepare for audiolm to be able to reconstruct from only coarse quanti…

0.10.10

commit loss was reduced over batch already

0.10.9

fix quantize dropout logic in rq

0.10.8

support masking, for phenaki

0.10.7

support masking, for phenaki

0.10.6

0.10.6

0.10.5

add quantizer dropout, for soundstream training

0.10.4

fix a "Tensor need to be contiguous" error in nccl distributed training