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/vit-pytorch
. Pull mirroring updated
Sep 19, 2024
.
0.29.1
9cd56ff2
·
CCT allow for rectangular images
·
Mar 26, 2022
0.29.0
9bae62f6
·
add proposed parallel vit from facebook ai for exploration purposes
·
Mar 23, 2022
0.28.2
c2b2db2a
·
fix window size of none for scalable vit for rectangular images
·
Mar 22, 2022
0.28.1
719048d1
·
some better defaults for scalable vit
·
Mar 22, 2022
0.28.0
d27721a8
·
add scalable vit, from bytedance AI
·
Mar 22, 2022
0.27.1
cb22cbbd
·
update to einops 0.4, which is torchscript jit friendly
·
Mar 22, 2022
0.27.0
6db20deb
·
add patch merger
·
Mar 01, 2022
0.26.7
1bae5d3c
·
allow for rectangular images for efficient adapter
·
Jan 31, 2022
0.26.6
25b38429
·
return None from extractor if no attention layers
·
Jan 28, 2022
0.26.5
64a07f50
·
epsilon should be inside square root
·
Jan 24, 2022
0.26.4
c1528acd
·
fix feature maps in Nest, thanks to @MarkYangjiayi
·
Jan 22, 2022
0.26.3
4dddaf60
·
decoder positional embedding needs to be reapplied...
·
Jan 06, 2022
0.26.2
28eaba61
·
0.26.2
·
Jan 03, 2022
0.26.1
91ed7387
·
0.26.1
·
Dec 30, 2021
0.26.0
fd16cbdf
·
add ViT for small datasets
https://arxiv.org/abs/2112.13492
·
Dec 28, 2021
0.25.6
e52ac419
·
allow extractor to only return embeddings, to ready for vision transformers to be used in x-clip
·
Dec 25, 2021
0.25.5
08918854
·
include tests in package for conda
·
Dec 22, 2021
0.25.3
976f4892
·
add some tests
·
Dec 22, 2021
0.25.2
563f2388
·
add some tests
·
Dec 22, 2021
0.25.1
2c368d1d
·
add extractor wrapper
·
Dec 21, 2021
Prev
1
2
3
4
5
6
7
8
9
…
11
Next