WIll-Xu35

@shyamala-methy go to the original insightface git and search for 'caffe', you will be able to find a download link.

@xiakj https://drive.google.com/open?id=1H2vAWU2Mappm9yY190WX_KxGVgE9RiK8

@xiakj Sorry, check this link: https://drive.google.com/drive/folders/1D6gG2WRvOvUI7NyJLInGl2zBD6zDPcoI

Hi all, I met the same problem here. I only modified the gpu number and the rest of the codes are not touched. It seems that the code is combining...

@twmht if your input shape is fixed and both width and height are multiples of 160, then crop is not needed.

The original test.py detects 43 faces. After setting the nms_threshold to 0.01, it still returns 41 faces. Seems like the nms function is somehow not working properly. And one more...

After downgrading the mxnet version in the docker to 1.3.0, the inference result are exactly the same. But why?

@qianchenghao Nope, I used dlav0 instead

+1