how to conduct cross-set erasing-inpainting (CSEI)?
I observed a related file named "create_inpaint_data_mini.py", but there was something wrong when I tried to run it. "from torchFewShot.models.models_gnn import create_models from src.config import Config from src.edge_connect_few_shot torchFewShot.models.net_similary" It seems that net_similarity&models_gnn in the folder "torchFewShot/models" and the folder "src" is missing.
Could you please upload the whole project which is reproducible?
To run this code, you need to first run a standard classification model on the training set, and then use "create_inpaint_data_mini.py" with edegeconect for inpaintng. Because this part of the code is only used for data preprocessing and does not involve a specific training and testing process, I did not release the complete code about create dataset. In this project, I gave the miniImagenet data set generated by this code. You can skip this step.