Setup ezpc repo by following instructions from here. Load the environment and install additional required packages:
source path/to/EzPC/mpc_venv/bin/activate
pip install mscviplib Pillow
python path/to/EzPC/Athos/CompileTFGraph.py --config config.json --role server
python path/to/EzPC/Athos/CompileTFGraph.py --config config.json --role client
python pre_process.py covid19positive.jpeg
This dumps input image as a numpy array (xray.npy)
python path/to/EzPC/Athos/CompilerScripts/convert_np_to_fixedpt.py --inp xray.npy --config config.json
This dumps xray_fixedpt_scale_12.inp.
Server on one machine:
./model_SCI_OT.out r=1 p=12345 < model_input_weights_fixedpt_scale_12.inp
Client on another machine:
./model_SCI_OT.out r=2 ip=123.231.231.123 p=12345 < xray_fixedpt_scale_12.inp > output.txt
You can change ip to 127.0.0.1 if both run on the same machine. (Will take around ~30 mins to run with a peak memory usage of ~8GB)
python path/to/EzPC/Athos/CompilerScripts/get_output.py output.txt config.json
Dumps model_output.npy as a flattened numpy array(1-D).
Run input graph with tensorflow.
python run_tf.py xray.npy
Dumps output in tensorflow_output.npy. Compare with output generated in 2PC computation:
python path/to/EzPC/Athos/CompilerScripts/comparison_scripts/compare_np_arrs.py tensorflow_output.npy model_output.npy
We get output as
Arrays matched upto 1 decimal points
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent