Swin Transformer 模型在 ImageNet-21k（1400 万张图像，21841 个类）上以 384x384 的分辨率进行预训练。

Go to file

YYJ-aaaa 62c9f13d2b first commit		2024-12-26 11:07:18 +08:00
.gitattributes	Add .gitattributes	2024-12-26 10:59:37 +08:00
README.md	Initial commit	2024-12-26 10:59:36 +08:00
config.json	first commit	2024-12-26 11:07:18 +08:00
preprocessor_config.json	first commit	2024-12-26 11:07:18 +08:00
pytorch_model.bin	first commit	2024-12-26 11:07:18 +08:00
tf_model.h5	first commit	2024-12-26 11:07:18 +08:00

README.md

swin-base-patch4-window12-384-in22k_a14066591556694016394467

Swin Transformer 模型在 ImageNet-21k（1400 万张图像，21841 个类）上以 384x384 的分辨率进行预训练。