Published May 14, 2022
| Version v1
Physical object
Open
Weights of two ViT-L models
Creators
Description
- vitl_16_pt22k.pth contains the weights of a large vision transformer pretrained on ImageNet-22K using iBOT (source)
- vitl_16_ct.pth contains the weights of the same model finetuned for 35 epochs on 165k CT slices (4k patients from 7 public datasets)
Files
Files
(2.4 GB)
| Name | Size | Download all |
|---|---|---|
|
md5:23605d0bf9217815324ec8a2c89b6aae
|
1.2 GB | Download |
|
md5:040f15dca745f3bc10997c32fe227d0b
|
1.2 GB | Download |