Arun
Arun
I think the molecules is not getting parsed properly. I made a small example to reproduce it: ``` from rdkit import Chem smiles = 'CCCN+(CCC)CCC.O=S(O)(=[OH+])c1cc(N=Nc2cc(S(=O)(O)=[OH+])c3cccnc3c2O)ccc1C=Cc1ccc(N=Nc2cc(S(=O)(O)=[OH+])c3cccnc3c2O)cc1S(=O)(O)=[OH+]' mol = Chem.MolFromSmiles(smiles) ``` The...
I think we are yet to update our docker images to include pytorch. Maybe @rbharath can clarify better here.
Hi @JdoesScience , thanks for reaching out. I hope you have gone through DeepChem tutorials. The tutorials gives an introduction to DeepChem. To fix errors in docs build, here are...
This tutorial was contributed by @ peastman. A git blame on the notebook will show the contributors to the tutorial.
Hey, thanks for your contribution. Added a couple of comments, looking forward for the updated PR.
Flow Network based Generative Models for Non-Iterative Diverse Candidate Generation (https://arxiv.org/abs/2106.04399, https://github.com/bengioe/gflownet) paper proposes an approach for tasks like molecule synthesis. Maybe this can be added to the list.
@rbharath I am not sure that the tutorial https://github.com/deepchem/deepchem/blob/master/examples/tutorials/Creating_Models_with_TensorFlow_and_PyTorch.ipynb is not much useful. We probably need a better tutorial on how to compose and use layers.
Also, most graph models in DeepChem takes `DGLGraph` as input. Having `dc.data.GraphDataset` can allow for more customization as well as compatibility with other formats like `torch_geometric.data.Data` and `DGLGraph`.
Yes, that will be great @neerajanand-coder . After discussion from weekly dev-call, we can start by moving `dc.feat.GraphData` to `dc.data.GraphData`. Regarding creation of `dc.data.GraphDataset`, we need to think of what...
Now it makes sense on why we are having it in `feat`. Thanks for the explanation! I have two more feature requests regarding `GraphData`: - When we print a `GraphData`...