Convert all weights of Caffe network to half precision floating point

cv.Net.shrinkCaffeModel(src, dst)
cv.Net.shrinkCaffeModel(..., 'OptionName',optionValue, ...)

Input

• src Path to origin model from Caffe framework contains single precision floating point weights (usually has .caffemodel extension).
• dst Path to destination model with updated weights.

Options

• LayersTypes Set of layers types which parameters will be converted. By default (not set), converts only Convolutional and Fully-Connected layers' weights, i.e {'Convolution', 'InnerProduct'}.

Note: Shrinked model has no origin float32 weights so it can't be used in origin Caffe framework anymore. However the structure of data is taken from NVidia's <https://github.com/NVIDIA/caffe Caffe fork>. So the resulting model may be used there.