google-deepmind / google-deepmind/deepmind-research
Geometry optimisation with DM21 in PySCF?
Open
- Dominant language
- Jupyter Notebook
- Stars
- 15.2k
- Forks
- 2.9k
- PR merge metrics
- No merged PRs in 30d
Description
Is it available to use DM21 for geometry optimization of molecule using PySCF?
It works only with HF nuclear gradients like as:
```
mf = scf.RKS(mol)
mf._numint = dm21.NeuralNumInt(dm21.Functional.DM21)
mf.apply(grad.rhf.Gradients).as_scanner()
```
But doesn't work with RKS gradients in pyscf.grad.rks.Gradients
AttributeError: 'NoneType' object has no attribute 'mol' in neural_numint.py in eval_xc()
How can I use RKS nuclear gradients?
Contributor guide
Assessment
This issue has not been assessed yet.