GitLab

* Support jit hessian

* fix

* fix

* Support  periodic table Indexing in builtin models

* flake8

* more

* fix

* fix cuda

* Allow passing pbc and cell to sequential

* revert

* Simplify ase interface

* fix

* fix

* fix

* flake8

* Update ase.py

* fix

* Use namedtuple to improve API

* improve

* Use modern type annotations for utils.py

* update (#377)

* Remove python2 support from CI, _six, etc. (#370)

* Remove python2 step 1

* badge

* Delete _six.py

* Simplify triple_by_molecule (#368)

* Simplify triple_by_molecule

* fix

* fix

* fix

* Use @ syntax (#371)

* Use modern type annotations for nn.py (#373)

* Use modern type annotations for nn.py

* flake8

* Use modern type annotations for models.py (#374)

* Update utils.py

* Update utils.py

* Update utils.py

* Update utils.py

* remove float32 dtypes from comp6.py

* fix device type

* Update utils.py

* Update utils.py

fix typo!

* Update utils.py

if I train molecule like Oxygen which contains only one element, the self.self_energies is a scalar and will make the operation self_energies = self.self_energies[species] crash. Besides, the species of Oxygen is [[3,3]] and is over the boundary of the energy list, so need to minus the minimum.

* Update utils.py

* enable EnergyShifter scripting

* fix

* fix

* Support len in ChemicalSymbolsToInts

* fix

* flake8

* Support 0 molecule subsets when loading dataset

* fix

* fix

* using least-squares to compute atomic self energies from the dataset

* self atomic energy calculation in the example training file

Add AdamW optimizer, Add docs of nnp training to train a real ANI-1x model, remove h5py and ignite from dependencies, rename MAEMetric -> MaxAEMetric, use torch.utils.tensorboard to replace tensorboardX (#224)

Fixes: https://github.com/aiqm/torchani/issues/183