README.md

Split_DNN

After dividing the VGG16 model by each layer, interence is performed separately on the edge server and client.

• edge_server_main.py: code for inference on an edge server
• client_vgg_main.py: code for inference on an edge device
• communication.py: code for communication between the edge server and the edge device
• Golden_Retriever_Hund_Dog.jpg: one image file to perform inference on
• imagenet_class_index.json: the indexes of imagenet
• models/vgg16.py: pretrained VGG16 model

=How to run=

• Write the IP address of the edge server in the clientCommunication() function in the client_vgg_main.py.
• The default IP address is the local address.
• python3 edge_server_main.py
• python3 client_vgg_main.py