honghuCode/mobileFacenet-ncnn

Issues

• 我在转模型到NPU设备上运行，请问模型的mean_value和std_value分别是什么呢

  #14 opened 2 years ago by shivabohemian
  1

• 为什么mobilefacenet提取的任意图片的特征向量都是一样的，下面是我的主程序

  #5 opened 6 years ago by songyy137222
  6

• face comparation poor results

  #11 opened 5 years ago by lucasjinreal
  1

• 从mxnet到caffe再到ncnn,转换过程很顺利没报错，但是统一图片的测试结果相差很大，是否哪里存在bug?

  #13 opened 5 years ago by Royzon
  1

• getFeatByMobileFaceNetNCNN() 函数

  #12 opened 5 years ago by Yang-zhizhang
  0

• 提取出来的特征之间的距离，同人和不同人之间的距离都很接近，是哪一步弄错了嘛

  #4 opened 5 years ago by glunce
  2

• the value of trp@fpr=0 is descend too much after i transfer the mxnet to caffe model

  #10 opened 6 years ago by lmmcc
  1

• mergebn needed?

  #9 opened 6 years ago by Royzon
  1

• param is too old, please regenerate

  #8 opened 6 years ago by af258963
  0

• how do you get mobilefacenet feature?

  #1 opened 6 years ago by Linxincode
  4

• caffe 训练的 文件 有？

  #7 opened 6 years ago by sunjunlishi
  0

• can you offer Cmakelists.txt? so i can test it in osx and linux

  #6 opened 6 years ago by haoweiliang1996
  0

• About ncnn.lib,ncnn_mobilefaceNet.lib and the mtcnn_ncnn.dll

  #2 opened 6 years ago by goxiaobu
  1

• 编译调试

  #3 opened 6 years ago by Andrawzyf
  1