To create the environment:
conda create --name fast_explore python==3.9
conda activate fast_explore
conda install -c conda-forge rdkit=2020.09.1
conda install pytorch cudatoolkit=11.1 -c pytorch -c nvidia
conda install -c conda-forge jupyterlab
conda install -c rdkit rdkit
conda activate mol_explore_3
conda install -c conda-forge hydra
conda install -c anaconda networkx
conda install -c conda-forge tensorboardx
conda install -c conda-forge cairosvg
pip install chemprop
pip install oddt
pip install -U scikit-learn==0.22.1
pip install p_tqdm
pip install selfies
pip install hydra-core --upgrade
pip install tensorboard
pip install -e . \
To run VQ-VAE pretrain (make sure to unzip the chembl data files):
python pretrain.py --data_dir data/chembl --vocab_path data/chembl/vocab_selfies.pl --model_type vqvae --output_dir output --batch_size 32 --latent_size 10 --n_embed 10 --depth 4 --hidden_size 200 --save_steps 5000 --autoregressive --vq_coef 1. --commit_coef 1.
To run RL search model:
python rl/train.py expname=gsk+jnk+qed+sa task=gsk+jnk+qed+sa init_basis=data/gsk3_jnk3_qed_sa/rationales.json hydra.run.dir=path_to_output_dir verbose=True env.sim_threshold=[.3,.4] env.sim_func=mixed frag_vae.model_path pretrain_models/chembl_pretrained