Welcome to Compressors’s documentation!

Why Compressors?

Compressors provides many ways to compress your model. You can use it either for CV either for NLP tasks.

Library separated into three parts:

• Distillation

• Pruning

• Quantization

There are two ways to use Compressors: with Catalyst or just use functional API.

Contents:

• Distillation
  • Philosophy and main idea
  • Submodules
    • Loss functions
    • Callbacks
    • Runners
    • Data
• Quantization
• Pruning
  • Runners
  • Callbacks
  • Utils
• Runners
• Models
  • Computer Vision
• Examples
  • Minimal Example
  • Minimal Complex Example
  • NLP
    • Huggingface transformers
  • CV