Chgnet

Changes
* Link to [Video tutorial](https://www.youtube.com/watch?v=Lm148F_1Dn4&feature=youtu.be) BowenD-UCB
* Allow setting MD start temperature BowenD-UCB
* Fixed bug triggerred by dtype janosh in https://github.com/CederGroupHub/chgnet/pull/95

**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/V0.3.1...v0.3.2

Hot fix release for v0.3.0 which is deprecated due to missing pretrained checkpoints (fixed in https://github.com/CederGroupHub/chgnet/pull/86/commits/0f0278649ed1d9d07ccfcef2aa11caacda5848fe). See https://github.com/CederGroupHub/chgnet/pull/86 for details.

🛠 Enhancements
* Add `CHGNet.version` property by janosh in https://github.com/CederGroupHub/chgnet/pull/86


**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/V0.3.0...v0.3.1

We release the most recent pretrained model: CHGNet 0.3.0 :tada: ([see details](https://github.com/CederGroupHub/chgnet/tree/main/chgnet/pretrained/0.3.0))
`CHGNet.load()` now defaults to '0.3.0' version.

Major changes:

1. Increased AtomGraph cutoff to 6A 55
2. Resolved discontinuity issue when no BondGraph presents 79
3. Added some normalization layers
4. Slight improvements on energy, force, stress accuracies

**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/v0.2.2...v0.3.0

Important Notes

v0.2.2 has a breaking change in model initialization from scratch i.e. `chgnet = CHGNet()`, to address 79.
The developer team is running another round of pretraining and the updated weights will be released in the next release.

For now, `chgnet = CHGNet.load()` loads same pretrained model as previous releases.
If user wants to load their own fine-tuned weights trained from previous releases, please do
`chgnet = CHGNet.from_file("your_weights_path", mlp_out_bias=True)`

What's Changed

- Run `cibuildwheel` on Apple silicon to publish pre-compiled macOS ARM binaries by janosh in https://github.com/CederGroupHub/chgnet/pull/78
- always report cov by lbluque in https://github.com/CederGroupHub/chgnet/pull/77
- 1.Dropped torch.det for MPS compatibility 2.better testing for rotation/supercell invariance by BowenD-UCB in [`4cee9e6`](https://github.com/CederGroupHub/chgnet/commit/4cee9e6aaa8fbe27a6ab9fd583d797e234cd3280)
- Added function to output site energies by BowenD-UCB in [`77845dc`](https://github.com/CederGroupHub/chgnet/commit/77845dc85ca7c5f2411a231eb8ad446a03e5f805)
- Added examples for parsing DFT by BowenD-UCB in [`2ebc57f`](https://github.com/CederGroupHub/chgnet/commit/2ebc57f92613803ee11fc7f9ce57ad7212a85e14)
- Fixed bug in 79, while not breaking loading pretrained weights by BowenD-UCB in [`1a558ea`](https://github.com/CederGroupHub/chgnet/commit/1a558ea4b96eac747fb27017bddcba161b7b8707)
- CHGNet now supports Apple MPS:tada:, pytorch-MPS has been tested with torch-2.0.1 by BowenD-UCB in [`516c422`](https://github.com/CederGroupHub/chgnet/commit/516c422f3bb8ee4d0dd75459ef4fa4ccd61baef0)
- added num_isolated_atoms property by BowenD-UCB in [`20ebc58`](https://github.com/CederGroupHub/chgnet/commit/20ebc587a806c93c289fc88be9bc4331dcd9dd05)
- allow specifying decay fraction in CosineAnnealing LR scheduler by BowenD-UCB in [`dbb0305`](https://github.com/CederGroupHub/chgnet/commit/dbb030571b85ad1d254adcd728ac2ebf9126c058)
- Moved all torch.int64 to torch.int32 for MPS support by BowenD-UCB in [`828478a`](https://github.com/CederGroupHub/chgnet/commit/828478a2645c90c99436ef8cc1c99ca2f470d7c3)
- added notes to avoid Colab bug by BowenD-UCB in [`a760ec8`](https://github.com/CederGroupHub/chgnet/commit/a760ec87fc0362de1caba7769a4029e6e5ece903)
- added installation from source by BowenD-UCB in [`030d897`](https://github.com/CederGroupHub/chgnet/commit/030d897b164c88c7a7a5778d0bb239ce56746e82)
- removed print message by BowenD-UCB in [`38a5cc4`](https://github.com/CederGroupHub/chgnet/commit/38a5cc44b001a1c440fa8eda990d2f28f04b7dc9)
- minor fix by BowenD-UCB in [`465f833`](https://github.com/CederGroupHub/chgnet/commit/465f833153b059d4f383f0b2bbf864b0cc3af729)
- fixed test by BowenD-UCB in [`d5ba24b`](https://github.com/CederGroupHub/chgnet/commit/d5ba24b9f265954bfb712a42e1795543b49af27e)
- loosen numerical tolerance by BowenD-UCB in [`b59d301`](https://github.com/CederGroupHub/chgnet/commit/b59d3013e62886b61de4946f384f5cf34c250561)
- added gMLP_norm option if gMLP is used at read_out by BowenD-UCB in [`195d48c`](https://github.com/CederGroupHub/chgnet/commit/195d48c62d14dcd3aafc4808d9f3b071bcec3222)


**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/v0.2.1...v0.2.2

More Supports for MD: LAMMPS and ASE-MD :zap:

Main updates in this release:

1. Implemented support for Nose-Hoover thermostat with ASE MD, choice of thermostat is now an option. BowenD-UCB tsihyoung
2. MD and structure optimization now defaults to CUDA with most available memory. Device selection is mostly automatic now. BowenD-UCB
3. MD now has option to generate hidden space features, allowing opportunity for post analysis. AegisIK BowenD-UCB
4. LAMMPS interface has been implemented, more tests to go. 57 advancesoftcorp
5. NVE MD is implemented with ASE-MD. BowenD-UCB
6. Isolated atom now defaults to warning in MD and relaxer (previously raised an error). Makes MD more stable BowenD-UCB
7. VASP parser now checks electronic convergence at every ionic step. BowenD-UCB
8. Updated paper reference and citation. janosh
9. tests and minor bug-fix. janosh BowenD-UCB
10. Publish pre-compiled wheels for all platforms to PyPI. janosh

New Contributors
* tsihyoung
* advancesoftcorp

**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/v0.2.0...v0.2.1

CHGNet is much faster now :rocket: :rocket: :rocket:

Main updates in this release:

1. Implement fast Cython graph converter in https://github.com/CederGroupHub/chgnet/pull/40 by AegisIK, yielding a 2 - 3x speed up in graph conversion depending on system size. Molecular dynamics now runs 1.5 - 2 x faster as well!

2. CHGNet API docs went live at https://chgnet.lbl.gov :tada: in https://github.com/CederGroupHub/chgnet/pull/39 by janosh

4. Better tests and docstrings by BowenD-UCB

All PRs since previous release
* API Docs by janosh in https://github.com/CederGroupHub/chgnet/pull/39
* Fix `ValueError: ndarray is not C-contiguous` CI errors by janosh in https://github.com/CederGroupHub/chgnet/pull/41
* 2.4x avg speedup on PMG struct to CHGNET graph conversion by AegisIK in https://github.com/CederGroupHub/chgnet/pull/40
* More descriptive names for single-letter variables by janosh in https://github.com/CederGroupHub/chgnet/pull/42
* Fix `ruff` `PERF203` + `PERF401` violations by janosh in https://github.com/CederGroupHub/chgnet/pull/43
* FIXED MEMORY LEAK in fast graph converter. Also cleaned up unneeded code. by AegisIK in https://github.com/CederGroupHub/chgnet/pull/44
* Add `CHGNet.todict` method needed for ASE JSON serialization by janosh in https://github.com/CederGroupHub/chgnet/pull/49
* Added garbage collection cyclic references turnoff for fast graph converter by AegisIK in https://github.com/CederGroupHub/chgnet/pull/50
* Solved final memory leak in fast converter by AegisIK in https://github.com/CederGroupHub/chgnet/pull/51
* Use `MaterialsProject2020Compatibility` to correct final energy in `fine_tuning.ipynb` by janosh in https://github.com/CederGroupHub/chgnet/pull/46

New Contributors
* AegisIK made their first contribution in https://github.com/CederGroupHub/chgnet/pull/40

**Full Changelog**: https://github.com/CederGroupHub/chgnet/compare/v0.1.04...v0.2.0