Auto-differentiable embedding of Physics and Torch Machine Learning (AdePT-ML):
This is a convienience library built on top of PyTorch to enable easy integration and training of hybrid models involving physics and deep learning modules.
Features
- Allows integration of torch.nn.module with numpy functions and enable training with torch optimizers.
- Pre-defined Modules and configs for physics and MLP architectures.
- Integrated training function with tensorboard support.
Installation
Installing with pip
pip install adeptml
Requirements (Automatically installed with pip):
- PyTorch (https://pytorch.org/)
- Joblib (https://joblib.readthedocs.io/en/latest/) (For loading and saving model parameters)
- Tensorboard
Usage:
The primary building block of this package is the Hybrid Model class. It neatly packages all the member models into one main Torch model and enables running forward inference as well as backpropagation. The class accepts as input an instance of the Hybrid Config class. This config is useful in defining all the constituent modules and their inputs.
As component modules, the Models module provides a straight forward MLP implementation as well as a Physics Module. This module is a torch Autograd wrapper which enables the integration of non-Torch numpy functions into a fully torch model and allows for training with torch optimizers.
API Documentation:
Visit our Read the docs page
Examples:
Refer to the tests file. Additional examples will be added soon.
1 maintainer
Metadata
1 maintainer
Release files for AdeptML 2.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| adeptml-2.1.tar.gz | 9.7 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| adeptml-2.1-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 20.7 kB
Release files / adeptml-2.1.tar.gz
| Download URL | adeptml-2.1.tar.gz |
|---|---|
| Size | 9.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
5047280b8a3c3e75cd217f051d182ca4bff501477405315a910e63928036c0e8
|
|
BLAKE2b-256 checksum How to use checksums |
58b324dee12562c4eefd01607d250aa7d77589dc55223edf30ae05d03fb882c4
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.4 CPython/3.12.7 Linux/6.6.87.2-microsoft-standard-WSL2
|
Release files / adeptml-2.1-py3-none-any.whl
| Download URL | adeptml-2.1-py3-none-any.whl |
|---|---|
| Size | 11.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
5621797a52980db335f9b982329dc54ff1e24fd737d01adeb1de72ac03a9892e
|
|
BLAKE2b-256 checksum How to use checksums |
003f9f59368155a1d01a65db301769cd974ed82ecee0f17c9e56fcad2266d005
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.4 CPython/3.12.7 Linux/6.6.87.2-microsoft-standard-WSL2
|
Security for AdeptML 2.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
