Source code for "Chemical Synthesis Prediction through Contrastive Molecular Representation Learning with Reaction Knowledge Graph"
We provide the datasets here for those interested in our paper:
The reaction datasets used in this study are stored in ./data/
ReaKE is mainly based on the following packages:
- python 3.9.16
- rdkit 2023.3.1
- numpy 1.24.3
- pytorch 1.12.1+cu116
- dgl 0.9.0
Also you can install the required packages follow there instructions (tested on a linux terminal):
conda env create -f environment.yaml
for pre-training:
cd src
python main.py --task pretrain --gnn tag
We provided the pre-trained GNN model in ./saved/, including GCN, GAT, GraphSAGE and TAG.
for reaction classification:
cd src
python main.py --task reaction_classification --pretrained_model tag_f00 --dataset schneider --batch_size 200
for yield prediction:
cd src
python main.py --task yield_prediction --pretrained_model tag_f00 --dataset buchwald_hartwig --batch_size 200
Citation:
Contact: In case you have questions, please contact Jiancong Xie: xiejc3@mail2.sysu.edu.cn