Releases: lucidrains/vector-quantize-pytorch
Releases · lucidrains/vector-quantize-pytorch
1.1.1
should never require gradients
1.1.0
add random projection quantizer, used by Google USM
1.0.7
cleanup
1.0.6
add the structured quantized dropout from encodec paper
1.0.5
add the structured quantized dropout from encodec paper
1.0.4
make sure accept_image_fmap works with residual vector quantization
1.0.2
make sure accept_image_fmap works with residual vector quantization
1.0.1
start using lerp for ema
1.0.0
handle only 1 embedding given, release 1.0.0, should be matured
0.10.15
fix a bug with replacement of stale codes in euclidean codebook, than…