James Betker d1007ccfe7 Adjustments to pixpro to allow training against networks with arbitrarily large structural latents ... - The pixpro latent now rescales the latent space instead of using a "coordinate vector", which **might** have performance implications. - The latent against which the pixel loss is computed can now be a small, randomly sampled patch out of the entire latent, allowing further memory/computational discounts. Since the loss computation does not have a receptive field, this should not alter the loss. - The instance projection size can now be separate from the pixel projection size. - PixContrast removed entirely. - ResUnet with full resolution added.		2021-01-12 09:17:45 -07:00
..
byol
fixup_resnet
flownet2@db2b7899ea
glean
pixel_level_contrastive_learning	Adjustments to pixpro to allow training against networks with arbitrarily large structural latents	2021-01-12 09:17:45 -07:00
srflow
styled_sr
stylegan
tecogan
transformers
vqvae	Support training imagenet classifier	2021-01-11 20:09:16 -07:00
__init__.py
arch_util.py
discriminator_vgg_arch.py
feature_arch.py
ResGen_arch.py
resnet_with_checkpointing.py	Support training imagenet classifier	2021-01-11 20:09:16 -07:00
RRDBNet_arch.py	Add injected noise into bypass maps	2021-01-07 16:31:12 -07:00
spinenet_arch.py
weighted_conv_resnet.py	Support training imagenet classifier	2021-01-11 20:09:16 -07:00