https://github.com/bytedance/ferminet_with_ecp
An implementation combining FermiNet with effective core potential (ecp). For paper, see https://arxiv.org/abs/2108.11661.
Science Score: 36.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (9.0%) to scientific vocabulary
Repository
An implementation combining FermiNet with effective core potential (ecp). For paper, see https://arxiv.org/abs/2108.11661.
Basic Info
- Host: GitHub
- Owner: bytedance
- License: apache-2.0
- Language: Python
- Default Branch: main
- Size: 23.4 KB
Statistics
- Stars: 15
- Watchers: 2
- Forks: 2
- Open Issues: 3
- Releases: 0
Metadata Files
README.md
Fermionic Neural Network with Effective Core Potential
An implementation combining FermiNet with effective core potential (ecp). For paper, see https://arxiv.org/abs/2108.11661.
This repository directly depends on FermiNet (Many thanks to this awesome software and the team behind it!). Certain files are modified from the corresponding ones in FermiNet, and we added comments prefixed with "MODIFICATION FROM FERMINET" on the introduced changes.
Installation
pip install -e . will install all required dependencies.
Usage
Workflow of ferminetecp is similar to the original FermiNet, which uses the ConfigDict from
ml_collections to configure the
system. A few example scripts are included under `ferminetecp/configs/ecp`.
ferminet_ecp --config ferminet_ecp/configs/ecp/X.py:Ga,1 --config.batch_size 256 --config.pretrain.iterations 100
To use ECP for atoms, define ECP-related fields in the corresponding config file. For instance,
```python
from pyscf import gto
from ferminetecp import baseconfig
def getconfig(inputstr): symbol, spin = inputstr.split(',') cfg = baseconfig.default() mol = gto.Mole() # Set up molecule mol.build( atom=f'{symbol} 0 0 0', basis={symbol: 'ccecpccpvdz'}, ecp={symbol: 'ccecp'}, spin=int(spin))
cfg.system.pyscf_mol = mol
return cfg
``` Moreover, we want to mention that we remove local energy outlier in training phase via config.optim.rm_outlier flag, which violates variational principle and needs to be turned off in inference phase.
Our experiments were carried out with jax==0.2.12 and jaxlib==0.1.65+cuda102. We hit some cuda issues with cuda 11, especially when training with KFAC.
Giving Credit
If you use this code in your work, please cite the associated paper.
@misc{li2021fermionic,
title={Fermionic Neural Network with Effective Core Potential},
author={Xiang Li and Cunwei Fan and Weiluo Ren and Ji Chen},
year={2021},
eprint={2108.11661},
archivePrefix={arXiv},
primaryClass={physics.chem-ph}
}
Owner
- Name: Bytedance Inc.
- Login: bytedance
- Kind: organization
- Location: Singapore
- Website: https://opensource.bytedance.com
- Twitter: ByteDanceOSS
- Repositories: 255
- Profile: https://github.com/bytedance
GitHub Events
Total
- Issues event: 1
Last Year
- Issues event: 1
Issues and Pull Requests
Last synced: about 1 year ago
All Time
- Total issues: 2
- Total pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Total issue authors: 2
- Total pull request authors: 0
- Average comments per issue: 2.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Past Year
- Issues: 0
- Pull requests: 0
- Average time to close issues: N/A
- Average time to close pull requests: N/A
- Issue authors: 0
- Pull request authors: 0
- Average comments per issue: 0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- KingDamLiu (1)
- menghaozhe (1)
- Aqsa-sk (1)