Skip to content
GitLab
Explore
Sign in
Tags
Tags give the ability to mark specific points in history as being important
This project is mirrored from
https://github.com/lucidrains/gigagan-pytorch
. Pull mirroring updated
Sep 19, 2024
.
0.1.24
77132d8d
·
final auxiliary loss wired up in gigagan trainer
·
Jul 22, 2023
0.1.23
7faa5d3a
·
add gradient penalty for vision aided discriminator
·
Jul 22, 2023
0.1.22
436b08d2
·
format
·
Jul 21, 2023
0.1.21
e43a2567
·
add preliminary support for vision aided discriminator as part of training
·
Jul 21, 2023
0.1.19
889b4b71
·
stick with convention of self attention before cross attention in upsampler
·
Jul 21, 2023
0.1.18
fb2a55e7
·
patch
·
Jul 21, 2023
0.1.17
211a2df3
·
also use bilinear upsample for rgb path in unet upsampler, following in stylegan design
·
Jul 21, 2023
0.1.16
1e396a76
·
default back to bilinear upsample for base generator as in stylegan
·
Jul 21, 2023
0.1.15
1d9160f7
·
for conditional training, complete the auxiliary clip contrastive loss
·
Jul 20, 2023
0.1.14
0ffcbad1
·
make sure vision-aided discriminator has same conditionality as generator and discriminator
·
Jul 20, 2023
0.1.12
60e261fb
·
prepare for vision aided discriminator, make sure it can be saved and loaded,...
·
Jul 20, 2023
0.1.11
11659dc5
·
Update version.py
·
Jul 20, 2023
0.1.10
3c656fb1
·
small prep for vision aided discr
·
Jul 20, 2023
0.1.9
4e805699
·
find_unused_parameters needs to be true
·
Jul 20, 2023
0.1.8
05c9668f
·
multi-gpu training on one machine works
·
Jul 20, 2023
0.1.7
1b4340f1
·
bring back the GLU activations in generator from lightweight gan
·
Jul 20, 2023
0.1.6
f0e01029
·
fix gradient penalty for mixed precision
·
Jul 20, 2023
0.1.5
ca3af02e
·
wait for everyone before doing ema step
·
Jul 20, 2023
0.1.4
cc52ba93
·
can pass in clip text encoding into generator and discriminator for conditional training
·
Jul 20, 2023
0.1.2
a9d14fed
·
complete mixed precision
·
Jul 20, 2023
Prev
1
2
3
4
5
6
7
Next