search

ORNL / HydraGNN

Distributed PyTorch implementation of multi-headed graph convolutional neural networks
BSD 3-Clause "New" or "Revised" License
68 stars 29 forks source link

Atomic descriptor torch smiles #177

Closed allaffa closed 1 year ago

allaffa commented 1 year ago

Looks alright to me. One minor thing is why need x.to(torch.float) in line 105 given x is already float? Maybe do the conversion for atomicdescriptors_torch_tensor

@pzhanggit this conversion was already in the older version of the routine. https://github.com/ORNL/HydraGNN/blob/c78e7ac5ecd79788b54c1f60253e57f6eaf9d738/hydragnn/utils/smiles_utils.py

I did not change this conversion in this PR only because this is not the object of the PR. :-) I made a small change in this regard. :)