Shi Yating
Shi Yating
@Jade999 我觉得是需要的,特征提取网络送入的应该是归一化好的人脸。 借此帖也想请教下博士 @hanson-young 。完成人脸检测后应该还有人脸归一化这一步,请问您了解到的NNIE有对应的仿射变换函数呢?谢谢!
谢谢。我在开发包中找到了hi_comm_svp.h,但是没有找到sample_comm_svp.h。请问您使用的sdk是什么版本?我是HiSVP_PC_V1.2.0.5。不知道是不是版本不一致呢?
@clancylian hi I open your shared pretrained mobilenet25 model link, https://pan.baidu.com/share/init?surl=P1ypO7VYUbNAezdvLm2m9w#list/path=%2F could you didn't give us the password?thank u
@xiakj @OPPOA113 我使用https://github.com/Charrin/RetinaFace-Cpp下方法转换后的R50,也是无法检测出来。请问二位解决这个问题了吗?能否发送我一份模型和prototxt,谢谢[email protected]
just build a vs project and add the cpp .cu files to the project.
make sure the param of crop layer crop_param { axis: 2 offset: 0 offset: 0 }
你的测试机是什么环境呢?显卡用的什么?
@Royzon Hi where did you download the mobilenet25 mxnet model?could you give me the website?thank u
have u trained resnet50 model trained on PASCAL VOC dataset?Which of the above two way to prepare training image do u choose?