DlibDotNet：针对Windows，MacOS和Linux用C ++和C＃编写的Dlib .NET包装器

# DNN Inception This program is ported by C# from examples\dnn_inception_ex.cpp. ## How to use? ## 1. Build 1. Open command prompt and change to &lt;DnnInception_dir&gt; 1. Type the following command ```` dotnet build -c Release ```` 2. Copy ***DlibDotNet.dll***, ***DlibDotNetNative.dll*** and ***DlibDotNetNativeDnn.dll*** to output directory; &lt;DnnInception_dir&gt;\bin\Release\netcoreapp2.0. **NOTE** - You should build ***DlibDotNetNative.dll*** and ***DlibDotNetNativeDnn.dll*** with CUDA. - If you want to run at Linux and MacOS, you should build the **DlibDotNet** at first. Please refer the [Tutorial for Linux](https://github.com/takuya-takeuchi/DlibDotNet/wiki/Tutorial-for-Linux) or [Tutorial for MacOS](https://github.com/takuya-takeuchi/DlibDotNet/wiki/Tutorial-for-MacOS). ## 2. Download demo data Download test data from the following urls. - http://yann.lecun.com/exdb/mnist/ - train-images-idx3-ubyte.gz - train-labels-idx1-ubyte.gz - t10k-images-idx3-ubyte.gz - t10k-labels-idx1-ubyte.gz And extract them and copy to extracted files to &lt;DnnInception_dir&gt;. ## 3. Run ```` cd <DnnInception_dir> dotnet run -c Release . The net has 43 layers in it. layer<0> loss_multiclass_log layer<1> fc (num_outputs=10) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<2> relu layer<3> fc (num_outputs=32) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<4> max_pool (nr=2, nc=2, stride_y=2, stride_x=2, padding_y=0, padding_x=0) layer<5> concat (1001,1002,1003) layer<6> tag1001 layer<7> relu layer<8> con (num_filters=4, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<9> skip1000 layer<10> tag1002 layer<11> relu layer<12> con (num_filters=4, nr=3, nc=3, stride_y=1, stride_x=1, padding_y=1, padding_x=1) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<13> skip1000 layer<14> tag1003 layer<15> relu layer<16> con (num_filters=4, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<17> max_pool (nr=3, nc=3, stride_y=1, stride_x=1, padding_y=1, padding_x=1) layer<18> tag1000 layer<19> max_pool (nr=2, nc=2, stride_y=2, stride_x=2, padding_y=0, padding_x=0) layer<20> concat (1001,1002,1003,1004) layer<21> tag1001 layer<22> relu layer<23> con (num_filters=10, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<24> skip1000 layer<25> tag1002 layer<26> relu layer<27> con (num_filters=10, nr=3, nc=3, stride_y=1, stride_x=1, padding_y=1, padding_x=1) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<28> relu layer<29> con (num_filters=16, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<30> skip1000 layer<31> tag1003 layer<32> relu layer<33> con (num_filters=10, nr=5, nc=5, stride_y=1, stride_x=1, padding_y=2, padding_x=2) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<34> relu layer<35> con (num_filters=16, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<36> skip1000 layer<37> tag1004 layer<38> relu layer<39> con (num_filters=10, nr=1, nc=1, stride_y=1, stride_x=1, padding_y=0, padding_x=0) learning_rate_mult=1 weight_decay_mult=1 bias_learning_rate_mult=1 bias_weight_decay_mult=0 layer<40> max_pool (nr=3, nc=3, stride_y=1, stride_x=1, padding_y=1, padding_x=1) layer<41> tag1000 layer<42> input<matrix> Traning NN... Epoch: 1 learning rate: 0.01 average loss: 0.480985 steps without apparent progress: 8 Epoch: 2 learning rate: 0.01 average loss: 0.105669 steps without apparent progress: 15 Epoch: 3 learning rate: 0.01 average loss: 0.0740526 steps without apparent progress: 318 Epoch: 4 learning rate: 0.01 average loss: 0.0604944 steps without apparent progress: 9 Saved state to inception_sync Epoch: 5 learning rate: 0.01 average loss: 0.0520312 steps without apparent progress: 326 Epoch: 6 learning rate: 0.01 average loss: 0.0459913 steps without apparent progress: 350 Epoch: 7 learning rate: 0.01 average loss: 0.0413119 steps without apparent progress: 10 Epoch: 8 learning rate: 0.01 average loss: 0.0379306 steps without apparent progress: 26 Saved state to inception_sync_ Epoch: 9 learning rate: 0.01 average loss: 0.0350026 steps without apparent progress: 542 Epoch: 10 learning rate: 0.01 average loss: 0.0325879 steps without apparent progress: 11 Epoch: 11 learning rate: 0.01 average loss: 0.0306371 steps without apparent progress: 550 Epoch: 12 learning rate: 0.01 average loss: 0.0287543 steps without apparent progress: 517 Epoch: 13 learning rate: 0.01 average loss: 0.0271211 steps without apparent progress: 12 Saved state to inception_sync Epoch: 14 learning rate: 0.01 average loss: 0.0258299 steps without apparent progress: 540 Epoch: 15 learning rate: 0.01 average loss: 0.0246587 steps without apparent progress: 518 Epoch: 16 learning rate: 0.01 average loss: 0.0235673 steps without apparent progress: 522 Epoch: 17 learning rate: 0.01 average loss: 0.0224451 steps without apparent progress: 558 Epoch: 18 learning rate: 0.01 average loss: 0.0216577 steps without apparent progress: 545 Saved state to inception_sync_ Epoch: 19 learning rate: 0.01 average loss: 0.0207857 steps without apparent progress: 566 Epoch: 20 learning rate: 0.01 average loss: 0.0197219 steps without apparent progress: 979 Epoch: 21 learning rate: 0.01 average loss: 0.0189983 steps without apparent progress: 989 Epoch: 22 learning rate: 0.01 average loss: 0.0183556 steps without apparent progress: 1037 Saved state to inception_sync Epoch: 23 learning rate: 0.01 average loss: 0.0176799 steps without apparent progress: 986 Epoch: 24 learning rate: 0.01 average loss: 0.0169129 steps without apparent progress: 970 Epoch: 25 learning rate: 0.01 average loss: 0.0161669 steps without apparent progress: 979 Epoch: 26 learning rate: 0.01 average loss: 0.0154576 steps without apparent progress: 983 Epoch: 27 learning rate: 0.01 average loss: 0.0146988 steps without apparent progress: 985 Saved state to inception_sync_ Epoch: 28 learning rate: 0.01 average loss: 0.0141198 steps without apparent progress: 984 Epoch: 29 learning rate: 0.01 average loss: 0.0135091 steps without apparent progress: 984 Epoch: 30 learning rate: 0.01 average loss: 0.0130994 steps without apparent progress: 991 Epoch: 31 learning rate: 0.01 average loss: 0.0125059 steps without apparent progress: 979 Epoch: 32 learning rate: 0.01 average loss: 0.012067 steps without apparent progress: 983 Saved state to inception_sync Epoch: 33 learning rate: 0.01 average loss: 0.0114886 steps without apparent progress: 998 Epoch: 34 learning rate: 0.01 average loss: 0.0109174 steps without apparent progress: 560 Epoch: 35 learning rate: 0.01 average loss: 0.0106421 steps without apparent progress: 986 Epoch: 36 learning rate: 0.01 average loss: 0.0103991 steps without apparent progress: 1060 Saved state to inception_sync_ Epoch: 37 learning rate: 0.01 average loss: 0.0100768 steps without apparent progress: 1061 Epoch: 38 learnin