Published February 16, 2025 | Version v1
Dataset Open

A Deep Learning-Augmented Density Functional Framework for Reaction Modeling with Chemical Accuracy

Creators

Description

# A Deep Learning-Augmented Density Functional Framework for Reaction Modeling with Chemical Accuracy

this manual will tell you how to train and test with you datasets.
## Install Requirements
Install package version
```python
python==3.9.0
pyscf==2.2.1
torch==2.0.0
ruamel.yaml==0.17.21
numpy==1.24.2
scipy==1.10.1
paramiko==3.1.0
```
Then we will install package [DeePKS-kit](https://github.com/deepmodeling/deepks-kit).
DeePKS-kit is a pure python library so it can be installed following the standard `git clone` then `pip install` procedure. Note that the two main requirements `pytorch` and `pyscf` will not be installed automatically so you will need to install them manually in advance. Below is a more detailed instruction that includes installing the required libraries in the environment.

We use `mamba` here as an example. So first you may need to install [Miniforge](https://github.com/conda-forge/miniforge). and install the requirement package.
```
mamba create -n deepks python=3.9.0 paramiko=3.1.0 numpy scipy=1.10.1 h5py ruamel.yaml=0.17.21 paramiko=3.1.0
mamba activate deepks
pip install torch==2.0.0 torchvision==0.15.1 torchaudio==2.0.1
pip install pyscf=2.2.1
```
Then you can install deepks
```
git clone https://github.com/deepmodeling/deepks-kit
cd deepks/
python setup.py install
```
or you can install
```
pip install git+https://github.com/deepmodeling/deepks-kit/
```

## Train our datasets
The project is Below,  `QM` document contain the each type DFT training parameters`train_input.yaml` and datasets path such as `train.raw`, and `get_energy.py` will give you output energy.

`test_sets` contain the descriptor of test sets and reults

`validate_sets` contain the results

 ```python
projects
├── QM
│   ├── B3LYP
│   │   ├── GRAM
│   │   ├── GRAMandT1X
│   │   └── T1X
└── validate_sets
    └── WHG_BHRE
        └── result
 ```
### Train
we use the `DeePHF@B3LYP` as a example.
```
cd QM/B3LYP/GRAMandT1X
deepks train train_input.yaml  -d train.raw -t valid.raw -o model.out/model.pth > model.out/log.iter 2> model.out/err.iter
```
Test
```
## you can check the test.raw error
deepks test -m model.out/model.pth -d test.raw -o test_corr/test >L1L2.out

## or you can use `get_energy.py` get the output `energy`
python get_energy.py --raw you_test.raw
```

Files

Files (1.2 GB)

Name Size Download all
md5:b649b9818a454e6d561583715f579f96
1.2 GB Download