IBM Analog Hardware Acceleration Kit
v0.5.0

Get started

  • Installation
  • Advanced installation guide
  • Using the pytorch integration
  • Glossary

Analog AI Concepts

  • Analog AI
  • Analog AI Hardware
  • Advantages and Challenges

Using the Simulator

  • Using aihwkit Simulator
  • Using Experiments

Analog DNN Training

  • Specialized Update Algorithms
  • Analog Training Presets

Analog DNN Inference

  • Inference and PCM statistical model

Advanced Guides

  • aihwkit design
  • Development setup
  • Development conventions
  • Project roadmap
  • Changelog

References

  • API Reference
  • Paper References
IBM Analog Hardware Acceleration Kit
  • »
  • aihwkit.inference package
  • Edit on GitHub

aihwkit.inference package¶

aihwkit.inference.compensation

Compensation methods such as drift compensation during analog inference.

aihwkit.inference.converter

Converter of weight matrix values into conductance values and back for analog inference.

aihwkit.inference.noise

Noise models to apply to converted weight values during analog inference.

High level inference tools.

Subpackages¶

  • aihwkit.inference.noise package
    • Submodules
      • aihwkit.inference.noise.base module
      • aihwkit.inference.noise.pcm module
      • aihwkit.inference.noise.custom module
  • aihwkit.inference.converter package
    • Submodules
      • aihwkit.inference.converter.base module
      • aihwkit.inference.converter.conductance module
  • aihwkit.inference.compensation package
    • Submodules
      • aihwkit.inference.compensation.base module
      • aihwkit.inference.compensation.drift module

© Copyright 2020, 2021 IBM Research. Revision c737a520.

Built with Sphinx using a theme provided by Read the Docs.