aptt 0.1.0

No project description provided

APTT – Antons PyTorch Tools

APTT (Antons PyTorch Tools) is a modular, extensible deep learning framework designed to streamline training, evaluation, and experimentation using PyTorch Lightning. It supports a wide range of model architectures, loss functions, evaluation metrics, and training utilities—across both vision and audio domains.

Features

• ✅ Wide range of supported model types (YOLO, ResNet, RNNs, WaveNet, etc.)
• 🧩 Pluggable callbacks (TorchScript export, TensorRT optimization, t-SNE visualization, etc.)
• 🧠 Built-in continual learning and knowledge distillation
• ⚙️ Modular structure (Heads, Losses, Layers, Metrics, Callbacks, etc.)
• 📊 Embedding visualization & analysis tools
• 🗂️ Flexible dataset loaders for audio and image tasks
• 🧪 Unit tests and full documentation with Sphinx

Project Structure

aptt/
├── aptt/                  # Core source code (models, callbacks, utils, etc.)
├── tests/                 # Unit tests
├── docs/                  # Sphinx-based documentation
├── README.md              # This file
├── pyproject.toml         # Build system and dependencies
└── LICENSE                # License information

---

Installation

# Clone the repository
git clone https://github.com/your-user/aptt.git
cd aptt

# (Optional) Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # or venv\Scripts\activate on Windows

# (For Sphinx Documentation)
apt-get install libgraphviz-dev


# Install dependencies
uv install .

Quick Start Example

from aptt.lightning_base.trainer import APTTTrainer

trainer = APTTTrainer(config_path="config.yaml")
trainer.train()

Documentation

To build the documentation locally:

cd docs
make html

The HTML output will be located in docs/_build/html/index.html.

License

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