Here is the source code for the submission paper "Retrieving GNN Architecture for Collaborative Filtering ". Thank you for your interest in our work!

First we will construct meta-train database for both tasks.

bash space_for_cf/run_batch.sh

Then we will do the meta train and searching.

bash run.sh

The generate yaml file move to configs in space_for cf and rerun the configs based on space_for cf.

the meta feature extractor in space_for_cf/datasets/dataset_to_meta.py

Here is the commit readme version of conference. We will release a more detail readme soon.