deepinsight / deepinsight/insightface
About Mixed Precision in ViT.
- Dominant language
- Python
- Stars
- 29.7k
- Forks
- 6.1k
- PR merge metrics
- No merged PRs in 30d
Description
Hi, I found some engineering tricks in your ViT backbone code([ViT](https://github.com/deepinsight/insightface/blob/master/recognition/arcface_torch/backbones/vit.py)).
in Attention Module
in Block Module
in forward_features of ViT
Could you explain the reason of using these codes? Thanks.
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.