Implementation for our paper, accpeted to MLDD workshop in ICLR 2022.

This is a minimum working version of the code used for the paper.

We upload a small version of binding affinity dataset originally from BindingDB. The dataset size is limited for the simple test of our code, so the test performance is not same with the paper.

conda env create --file environment.yaml
conda activate metadta

The simple model training code is

python train.py --use_latent_path 

The use_latent_path option is the option for the latent embedding path, which is from the Attentive Neural Process