I've got some errors when pretraining.

[Syzseisus]

Hi, thank you for your great work and cool code.

I open this issue because of some errors when pretraining.

1. 'Mol' object has no attribute 'pSmi'. I can't find this error from Google.
   This error occurs at line: frags = list(BRICSDecompose(mol, returnmols=True)).
   example mol smiles:
   1. CC(=O)NC1C(O)OC(CO)C(OC2OC(CO)C(OC3OC(CO)C(OC4OC(CO)C(O)C(O)C4OC4OC(CO)C(OC5OC(CO)C(O)C(O)C5O)C(O)C4NC(C)=O)C(OC4OC(CO)C(OC5OC(CO)C(OC6OC(CO)C(O)C(O)C6O)C(O)C5NC(C)=O)C(O)C4OC4OC(CO)C(OC5OC(CO)C(O)C(O)C5O)C(O)C4NC(C)=O)C3O)C(O)C2NC(C)=O)C1O
   2. C=CCOC1C(OCc2ccccc2)C(OCc2ccccc2)C(OCc2ccccc2)C(OCc2ccccc2)C1OC1OC(COCc2ccccc2)C(OC2OC(COC3OC(COCc4ccccc4)C(OCc4ccccc4)C(OCc4ccccc4)C3OC3OC(COC(=O)CCl)C(OCc4ccccc4)C(OCc4ccccc4)C3OC3OC(COCc4ccccc4)C(OCc4ccccc4)C(OCc4ccccc4)C3OCc3ccccc3)C(OC3OC(COCc4ccccc4)C(OCc4ccccc4)C(OCc4ccccc4)C3NC(C)=O)C(OCc3ccccc3)C2OC(C)=O)C(OCc2ccccc2)C1N=[N+]=[N-]
2. Timeout error even though I set the threshold=10secs.
3. And OOM. I use batch size of 512 with data pubchem-10m-clean.txt.

What is your development environment?
Mine is an A6000 with 48GB VRAM.

Best,
Wooseong Cho.

[yuyangw]

Hi Wooseong,

Thanks for your interest in our work.

1. I haven't come across the issue before. Maybe try to sanitize the SMILES before using BRICS decomposition?
2. In that case, I would suspect that BRICS decomposition fails to break the molecular and thus gets stuck.
3. With 48GB, OOM could happen. We used a workstation with 128GB RAM.

Hope this helps.

Best,
Yuyang

[Syzseisus]

Hi,
I couldn't do it until now, but there is still issue.
I do exactly as it is written in README.md,

1. create environment same as README.md
2. clone iMolCLR repository
3. download data and unzip, rename ./data/pubchem-10m-clean.txt -> ./data/pubchem-10k.txt as in config.yaml.
4. Run python imolclr.py, I got a Timeout Error.
5. as you say, I sanitize the SMILES in dataset.py file like:

class MoleculeDataset(Dataset):
    def __init__(self, smiles_data):
        super(Dataset, self).__init__()
        self.smiles_data = smiles_data

    def __getitem__(self, index):
        mol = Chem.MolFromSmiles(self.smiles_data[index])
        Chem.SanitizeMol(mol)
        ...

6. Run, and again Timeout error.

what is the problem?