https://github.com/njzjz/dpdata_ani

dpdata plugin for ANI
https://github.com/njzjz/dpdata_ani

Last synced: 2 months ago
JSON representation

dpdata plugin for ANI

• Host: GitHub
• URL: https://github.com/njzjz/dpdata_ani
• Owner: njzjz
• License: lgpl-3.0
• Created: 2022-05-10T00:39:54.000Z (over 2 years ago)
• Default Branch: master
• Last Pushed: 2024-05-28T20:16:23.000Z (7 months ago)
• Last Synced: 2024-05-29T11:15:05.955Z (7 months ago)
• Language: Python
• Size: 399 KB
• Stars: 0
• Watchers: 2
• Forks: 1
• Open Issues: 3
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# dpdata_ani

The [dpdata](https://github.com/deepmodeling/dpdata) plugin for ANI.

## Features

- `ani/1x` format to load [ANI-1x dataset](https://doi.org/10.1038/s41597-020-0473-z)

- `ani/1x`, `ani/2x`, and `ani/1ccx` driver to predict using ANI models with automatic batch size

- `aiqm1` driver for [AIQM1](https://doi.org/10.1038/s41467-021-27340-2) which is based on ANI

## Installation

```sh

pip install dpdata-ani

```

To enable AIQM1,

```sh

pip install dpdata-ani[aiqm1]

```

## Usage

First, download ANI-1x dataset:

```sh

wget https://figshare.com/ndownloader/files/18112775 -O ani1x.hdf5

```

Load the ANI-1x dataset, and predict it with ANI-2x models:

```py

import dpdata

ani = dpdata.MultiSystems().from_ani_1x("ani1x.hdf5")

predict = ani.predict(driver="ani/2x")

predict.to_deepmd_hdf5("predict.hdf5")

```

The inferred data will be stored in `predict.hdf5`.