You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: README.md
+8Lines changed: 8 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -26,6 +26,14 @@
26
26
* The Hugging Face Hub (https://huggingface.co/timm) is now the primary source for `timm` weights. Model cards include link to papers, original source, license.
27
27
* Previous 0.6.x can be cloned from [0.6.x](https://github.com/rwightman/pytorch-image-models/tree/0.6.x) branch or installed via pip with version.
28
28
29
+
### Oct 20, 2023
30
+
*[SigLIP](https://huggingface.co/papers/2303.15343) image tower weights supported in `vision_transformer.py`.
31
+
* Great potential for fine-tune and downstream feature use.
32
+
* Experimental 'register' support in vit models as per [Vision Transformers Need Registers](https://huggingface.co/papers/2309.16588)
33
+
* Updated RepViT with new weight release. Thanks [wangao](https://github.com/jameslahm)
34
+
* Add patch resizing support (on pretrained weight load) to Swin models
35
+
* 0.9.8 release pending
36
+
29
37
### Sep 1, 2023
30
38
* TinyViT added by [SeeFun](https://github.com/seefun)
31
39
* Fix EfficientViT (MIT) to use torch.autocast so it works back to PT 1.10
0 commit comments