template-nn 0.1.4

A neural network model architecture template

Template NN

Template NN is a lightweight, easy-to-use template built to streamline neural network prototyping. At its core, it leverages the power of PyTorch.

Huge thanks to the PyTorch team for enabling projects like this.

Installation

[!WARNING] Breaking Changes in 0.1.4

This release includes directory renames, module restructuring, and function removals.
If you're upgrading from 0.1.3 or earlier, please read the Release Notes before updating.

You can install template-nn via pip:

pip install template-nn

Or clone the repository and install it locally:

git clone https://github.com/gabrielchoong/template-nn.git
cd template-nn
pip install -r requirements.txt
pip install .

Documentation

For detailed documentation, including usage instructions and examples, visit the online documentation at Documentation.

Feature Preview

import torch.nn as nn
from template_nn import FNN

# Single hidden layer model
model = FNN({
    "input_size": 10,
    "output_size": 5,
    "hidden_sizes": 5,
    "activation_functions": [nn.ReLU()]
})

# Deep network with 5 hidden layers
deep = FNN({
    "input_size": 10,
    "output_size": 5,
    "hidden_sizes": [8] * 5,
    "activation_functions": [nn.ReLU()] * 5
})

Releases and Contributing

This project is currently in its alpha stage. Please file an issue if you found a bug.

To read about the motive and direction of this project, see Roadmap. To read more about the current releases, see Release Notes.

Contributions are welcomed! Please see Contributions for information on contributing.

Contributors

This project is currently being developed by Gabriel.

License

This project is licensed under the MIT License - see the LICENSE file for details.