dnnc_miniGoogleNet.sh

#!/bin/bash

ML_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null && cd ../../.. && pwd )"
export ML_DIR
$(which dnnc &> /dev/null) || export PATH=$HOME/ML/DNNDK/tools:$PATH
#DNNDK_ROOT=$HOME/ML/DNNDK/tools

net=miniGoogleNet
work_dir=$ML_DIR/deephi/miniGoogleNet/quantiz
model_dir=${work_dir}/decent_output
output_dir=${work_dir}/dnnc_output

echo "Compiling network: ${net}"

dnnc --prototxt=${model_dir}/deploy.prototxt     \
       --caffemodel=${model_dir}/deploy.caffemodel \
       --output_dir=${output_dir}                  \
       --net_name=${net}                           \
       --dpu=4096FA                                \
       --cpu_arch=arm64                            \
       --mode=debug                                \
       --save_kernel



echo " copying dpu elf file into /../zcu102/baseline/model/arm64_4096 "
cp ${output_dir}/dpu_${net}\_*.elf  ${output_dir}/../../zcu102/baseline/model/arm64_4096

echo " copying the test images to be used by the ZCU102"
cp -r $ML_DIR/input/cifar10_jpg/test ${output_dir}/../../zcu102/test_images

#mv ${net}_kernel*.* ${work_dir}/dnnc_output

# The following lines are for documentation only. Leave them commented
# now you can create the tar archive that you will then scp/ssh copy from HOST PC to TARGET board 
## cd ~/ML/cifar10/deephi/miniVggNet/zcu102/baseline
## ln -s ../test_images ./test_images
## cd ..
## tar -xvf baseline.tar ./baseline