Metadata-Version: 2.4
Name: lema
Version: 0.0.1.dev2
Author-email: Xianghao Wang <wxianghao@outlook.com>
Requires-Python: >=3.12
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: loguru>=0.7.3
Requires-Dist: torch>=2.11.0
Requires-Dist: torchvision>=0.26.0
Dynamic: license-file

# LeMA: Multi-GPU Levenberg-Marquardt Optimizer Powered by NVIDIA cuPyNumeric and Legate

**LeMA** is a multi-GPU Levenberg-Marquardt optimizer for deep learning, implemented in PyTorch. The Levenberg-Marquardt algorithm is a second-order optimizer using the Jacobian matrix to compute the optimal parameter updates.

## Get Started

### 1. Install
```bash
pip install --prev lema
```

## Hands-on Examples
* Train convolutional neural network on MNIST dataset: [experiments/mnist](experiments/mnist/)
* TODO: Train physics-informed neural network
