tiny-dnn can import Caffe's trained models.

• Google protobuf
• OpenCV

1 Use protoc to generte caffe.pb.cc and caffe.pb.h.

cd tiny_dnn/io/caffe
protoc caffe.proto --cpp_out=./

2 Compile tiny_dnn/io/caffe/caffe.pb.cc and examples/caffe_converter/caffe_converter.cpp and link them.

./caffe_converter.bin [model-file] [trained-file] [mean-file] [label-file] [img-file]

In the pre-trained CaffeNet model,

./caffe_converter.bin\
 deploy.prototxt\
 bvlc_reference_caffenet.caffemodel\
 imagenet_mean.binaryproto\
 synset_words.txt\
 cat.jpg

• tiny-dnn's converter only supports single input/single output network without branch.