lightorch 0.0.1

Pytorch & Lightning based framework for research and ml-pipeline automation.

LighTorch

A Pytorch and Lightning based framework for research and ml pipeline automation.

Modules

Set useful architectures for several tasks.

• Fourier Convolution.
• Partial Convolution. (Optimized implementation)
• Grouped Query Attention, Multi Query Attention, Multi Head Attention. (Interpretative usage)
• Normalization methods.
• Positional encoding methods.
• Embedding methods.
• Useful criterions.
• Useful utilities.
• Built-in Default Feed Forward Networks.
• Adaptation for $\mathbb{C}$ modules.

Features

• Built in Module class for:
  • Adversarial training.
  • Supervised, Self-supervised training.
• Multi-Objective optimization and Hyperparameter tuning with optuna.
• Built-in default architectures: Transformers, VAEs, autoencoders for direct training on given data.

Contact

• Linkedin
• GitHub
• Email: [e-mail]