Alexander

@Oumoxiluoyi https://github.com/lippman1125/face-alignment-at-3000fps 我加了2进制模型的保存和载入的代码。

如果只训练12个特征点，模型只有6MB多。

I build a pytorch version for FAN, can be downloaded from link https://github.com/lippman1125/pytorch_FAN

Same question as above

@xuanyuyt https://github.com/lippman1125/faceboxes_lqy 实现了稠密化

你好，我的微信号：lippman1125

https://github.com/lippman1125/faceboxes_lqy 实现了整个测试评估流程

> Looks that it is very rarely seen. s_scale is a floating pointed number, it generally does not become 0 during training. I will try to add an eps, when...

@ysuwl please use commands from tutorials [https://github.com/meituan/YOLOv6/blob/main/tools/qat/README.md](url) to perform quantization

> Can you provide a code example on how you would want this to work? An example as follow, I want keep method 'get_length', but after compiled by using TRTorch,...