torchdrug
torchdrug copied to clipboard
DeepGraphLearning
Can't kekulize mol.
When I ran the code below（same as the demo in retrosynthesis tutorial）
from torchdrug import data, datasets, utils
from torchdrug.utils import plot
from torchdrug import core, models, tasks
synthon_dataset = datasets.USPTO50k("~/data", as_synthon=True,
node_feature="synthon_completion",
kekulize=True)
torch.manual_seed(1)
synthon_train, synthon_valid, synthon_test = synthon_dataset.split()
synthon_model = models.RGCN(input_dim=synthon_dataset.node_feature_dim,
hidden_dims=[256, 256, 256, 256, 256, 256],
num_relation=synthon_dataset.num_bond_type,
concat_hidden=True)
synthon_task = tasks.SynthonCompletion(synthon_model, feature=("graph",))
synthon_optimizer = torch.optim.Adam(synthon_task.parameters(), lr=1e-3)
synthon_solver = core.Engine(synthon_task, synthon_train, synthon_valid,
synthon_test, synthon_optimizer,
gpus=[0], batch_size=128)
synthon_solver.train(num_epoch=10)
synthon_solver.evaluate("valid")
synthon_solver.save("~/data/g2gs_synthon_model.pth")
an exception appeared:
Chem.Kekulize(mol)
rdkit.Chem.rdchem.KekulizeException: Can't kekulize mol. Unkekulized atoms: 11
Have anyone encountered that case, and how did you solve? Vary appreciated for relpies!
Sorry for the inconvience. I've changed rdkit version to 2021.9.2, which can address the problem in my case. Maybe you can have a try.
If it doesn't work for you, could you please tell us your pytorch and rdkit version? Since I'm not a developer for the retrosynthesis part, maybe @KiddoZhu and @shichence will help you on this.
Sorry for the inconvience. I've changed rdkit version to 2021.9.2, which can address the problem in my case. Maybe you can have a try.
If it doesn't work for you, could you please tell us your pytorch and rdkit version? Since I'm not a developer for the retrosynthesis part, maybe @KiddoZhu and @shichence will help you on this.
Thx! The problem is solved with the rdkit version you offered, the version I previously used is 2022.9.3 .
For anyone working on OSX (M1), I was able to succeed using version 2021.9.4.
pip install rdkit-pypi-2021.9.4
