yja1

in dataset.py : size = (uniform(-1.0, 1.0) * sigma + 1) * face_max_size when size is small,the crop_box width or height will be zero. then cause error on: crop_img_np =...

is there mobilenet0.25 pretrained model?

forward time and model size

1

I train a model: base cnn part use your mobilenet-ssd; next use SSH M1 M2 M3; change dwconv to DepthwiseConvolution use [https://github.com/yonghenglh6/DepthwiseConvolution](url) the final model size is 101M, (SSH vgg16...

 mbv3_small_75_light

   下边这里是怎么推导出来的呢？对应上边3种情况？ 

baseline

11

i also try this, the result is very bad. do you try table 5 in paper? I try, but the result of table 5 the first line only 68, far...

cmake .. -DUSE_MPI=ON -DCUDNN_INCLUDE=/cudnn5/include -DCUDNN_LIBRARY=/cudnn5/lib64/libcudnn.so -DBOOST_LIBRARY=/software/boost1.66.0/lib -DBOOST_ROOT=/software/boost1.66.0 -DGFLAGS_INCLUDE_DIR=/usr/include/gflags -DGFLAGS_LIBRARY=/usr/lib/x86_64-linux-gnu -DGLOG_INCLUDE_DIR=/usr/include/glog -DGLOG_LIBRARY=/usr/lib/x86_64-linux-gnu -DLMDB_INCLUDE_DIR=/usr/include -DLMDB_LIBRARIES=/usr/lib/x86_64-linux-gnu -DLevelDB_INCLUDE=/usr/include/leveldb -DLevelDB_LIBRARY=/usr/lib/x86_64-linux-gnu -DProtobuf_LIBRARY=/usr/lib/x86_64-linux-gnu ``` when make: the error: cudnn.hpp:108:70: error: too few arguments to function ‘cudnnStatus_t cudnnSetConvolution2dDescriptor(cudnnConvolutionDescriptor_t,...

test model error

5

run python test.py CenterNet-104 --testiter 480000 --split validation **error msg:** loading parameters at iteration: 480000 building neural network... module_file: models.CenterNet-104 total parameters: 210062960 loading parameters... loading model from cache/nnet/CenterNet-104/CenterNet-104_480000.pkl Traceback...

测试嘴以下是artifact

4

采用不同方式处理source img(face_alignment代码crop ; resize 到256x256;)得到的结果都是糊掉的

histogram_matching has cpu().numpy(), how about training time?