0.0.19

make sure commitment loss does not account for null vertex token. pla…

0.0.18

recon loss must account for variable lengthed faces

0.0.17

already checked for in vq library

0.0.16

works like a charm

0.0.15

able to derive face mask directly from codes based on padding id, for…

0.0.14

make sure readme runs

0.0.12

start handling variable lengthed faces

0.0.11

tokenizing from the autoencoder should always no grad and eval mode, …

0.0.10

can also train transformer off raw face data, if autoencoder has fini…

0.0.9

automatically take care of eos