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

0.0.8

fix can eos logic

0.0.7

axial embedding for each vertex of each face

0.0.6

some asserts

0.0.5

end to end generation

0.0.4

take care of the smoothing of the positions for autoencoder reconstru…

0.0.3

transformer forward pass

0.0.2

able to decode codes back to continuous coordinates (9) for each face…