Skip to content

Latest commit

 

History

History
31 lines (22 loc) · 1.04 KB

PRETRAIN.md

File metadata and controls

31 lines (22 loc) · 1.04 KB
Raw

Pretrain 3DMolMS on QM9

Please set up the environment as shown in step 0 from README.md.

Step 1: Download the QM9 dataset [here]. The structure of data directory is:

|- data
  |- qm9
    |- dsgdb9nsd.xyz.tar.bz2
    |- dsC7O2H10nsd.xyz.tar.bz2
    |- uncharacterized.txt

Step 2: Use the following commands to preprocess the datasets. The settings of datasets are in ./preprocess_etkdgv3.yml.

python ./src/qm92pkl.py --data_config_path ./src/molnetpack/config/preprocess_etkdgv3.yml

Step 3: Use the following commands to pretrain the model. The settings of model and training are in ./config/molnet_pre.yml.

python ./src/pretrain.py --train_data ./data/qm9_etkdgv3_train.pkl \
--test_data ./data/qm9_etkdgv3_test.pkl \
--model_config_path ./src/molnetpack/config/molnet_pre.yml \
--data_config_path ./src/molnetpack/config/preprocess_etkdgv3.yml \
--checkpoint_path ./check_point/molnet_pre_etkdgv3.pt