pengmiao-usc/GraphVision

Evaluating the feasibility of GNN algorithms with Image Classification.

GraphVision

Github repository for the Final project for CS GY 6953 Deep Learning

Our experiments are grouped into three categories and can be found in their respective directory:

• GENConv
• GeneralConv
• GATConv

Within each category, we listed notebooks corresponding to experiments arranged by

• the number of layers, such as 4, 8, 16, and 32
• the hidden dimension, such as 128 and 256
• learning rates, such as 0.01 (the default) and 0.001