Souvik Panda I'm delighted to hear you're attempting to identify embeddings of a drug dataset from MoleculeNet using graph autoencoders. However, it appears that you've been experiencing some difficulty with mistakes.

The first thing I would suggest is double-checking the input data to ensure it is in the right format. SMILES strings, which are a standard technique of describing chemical structures as strings, should be present in the CSV file you mentioned. The smile strings must be proper, and invalid smiles must be handled individually.

It's also critical to ensure that any pre-processing tasks, such as converting SMILES strings to graphs, are carried out appropriately. RDKit, a prominent open-source toolkit for cheminformatics, is one library you may utilize for this.

Regarding the Colab link you provided, it's impossible for me to assist without seeing the precise code and issues you're having. If you're still having problems, I recommend asking a query on a site like Stack Overflow and including the exact error message and traceback. You should also consider posting in the Moleculenet community or forums connected to cheminformatics or graph neural networks, where other academics working on comparable problems may be able to assist you.

It's also worth noting that there are multiple pre-trained models available on the MoleculeNet repository, so instead of training your own, you might try utilizing one of them as a starting point.

Finally, I recommend double-checking your code to ensure that you are using the right version of the packages and that your dataset is properly loaded. Small errors in the code can sometimes cause huge problems, and fixing them may solve the situation.

Qamar UI Ishlam can you personally contact with me if possible so that I can share you the errors.

Can anyone tell me the best way to find the embedding of a drug dataset using graph autoencoder of moleculenet dataset bcz i m facing many types of errors

DEAR Souvik Panda

Is your problem downloading the CSV file? or running embeddings of drug datasets? I didn't get it

running embeddings of drug datasets