Deep-Learning-Blocks

A library with customized PyTorch layers and model components.

• Install
• Documentation
• What's available
• Bug/Feature

Install:

To install the latest stable version:

pip install deepblocks

For a specific version:

pip install deepblocks==0.1.9

To install the latest, but unstable version:

pip install git+https://github.com/blurry-mood/Deep-Learning-Blocks

What's available:

Networks

• ConvMixer
• U-Net
• ICT-Net

Layers

• ConvMixer Layer
• Flip-Invariant Conv2d
• Squeeze-Excitation Block
• Dense Block
• Multi-Head Self-Attention
• Multi-Head Self-Attention V2

Activations

• Funnel ReLU

Loss Functions

• Focal Loss
• AUC Loss
• AUC Margin Loss
• KL Divergence Loss

Regularization functions

• Anti-Correlation

Self-supervised Learning

• Barlow Twin
• DINO

Optimizers

• SAM

Documentation:

The current documention is hosted here

Bug or Feature:

Deepblocks is a growing package. If you encounter a bug or would like to request a feature, please feel free to open an issue here.