This folder contains the molecule generation script. The polymer generation experiment in the paper can be reproduced through the following steps:
Extract substructure vocabulary from a given set of molecules (replace data/polymers/all.txt with your own molecules):
python get_vocab.py --min_frequency 100 --ncpu 8 < ../data/polymers/all.txt > vocab.txt
The --min_frequency means to discard any large motifs with lower than 100 occurances in the dataset. The discarded motifs will be decomposed into simple rings and bonds. Change --ncpu to specify the number of jobs for multiprocessing.
Preprocess the dataset using the vocabulary extracted from the first step:
python preprocess.py --train ../data/polymers/train.txt --vocab ../data/polymers/inter_vocab.txt --ncpu 8
mkdir train_processed
mv tensor* train_processed/
Train the generative model with KL regularization weight beta=0.1 and VAE latent dimension 24. You can change it by --beta and --latent_size argument.
mkdir -p ckpt/tmp
python vae_train.py --train train_processed/ --vocab ../data/polymers/inter_vocab.txt --save_dir ckpt/tmp
python sample.py --vocab ../data/polymers/inter_vocab.txt --model ckpt/inter-h250z24b0.1/model.19 > outputs.txt