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/video-diffusion-pytorch
. Pull mirroring updated
Sep 19, 2024
.
0.3.10
2a78aac0
·
make sure initial temporal attention gets rotary embeddings
·
May 09, 2022
0.3.9
21cd0376
·
bring in an initial temporal attention after the first initial convolution, to...
·
May 09, 2022
0.3.8
29014224
·
remove convnext blocks, as seeing first good results on moving mnist, thanks...
·
May 05, 2022
0.3.7
549f7b07
·
experiment tracker agnostic
·
May 04, 2022
0.3.6
9690e0aa
·
change focusing on present to be a probability, so that a single batch could...
·
May 04, 2022
0.3.5
3a0f1f2b
·
start using torch inference mode
·
May 04, 2022
0.3.4
4a0b6ee6
·
add additional rotary positional embedding for time attention
·
May 04, 2022
0.3.3
b645737b
·
allow for gradient clipping
·
May 04, 2022
0.3.2
8f3ecb15
·
allow for gradient clipping
·
May 04, 2022
0.3.1
2cc187fd
·
make sure some hyperparameters for resnet and convnext blocks are configurable
·
May 04, 2022
0.2.5
44ef9c21
·
offer old resnet blocks from original ddpm paper, in case convnexts are...
·
May 04, 2022
0.2.4
9daf6305
·
extra norm on the queries as well for linear attention, plus extra output...
·
May 04, 2022
0.2.3
233c1d69
·
offer way to turn off sparse linear attention
·
May 04, 2022
0.2.2
f3afff93
·
give an initial conv
·
May 01, 2022
0.2.1
06e1761a
·
0.2.1
·
May 01, 2022
0.2.0
84bf5354
·
keep time embedding dimension as dim x 4, as openai does
·
Apr 29, 2022
0.1.6
b68bd6b9
·
automate more
·
Apr 24, 2022
0.1.5
497b349e
·
offer way to load from last checkpoint, with `trainer.load(-1)`
·
Apr 22, 2022
0.1.2
f2f41b88
·
no classifier free guidance if unet is not conditioned
·
Apr 21, 2022
0.1.1
ee013c7e
·
small padding for the sampled grid of videos
·
Apr 21, 2022
Prev
1
2
3
Next