Data

Tools

Indexes

Applications

Experiments

Open Source Science

enpheeph

A modern Neural Fault Injection Framework

https://github.com/alexei95/enpheeph

Science Score: 36.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ✓
    Academic publication links
    Links to: arxiv.org
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (12.2%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

A modern Neural Fault Injection Framework

Basic Info
  • Host: GitHub
  • Owner: Alexei95
  • License: agpl-3.0
  • Language: Python
  • Default Branch: main
  • Homepage:
  • Size: 161 MB
Statistics
  • Stars: 2
  • Watchers: 0
  • Forks: 0
  • Open Issues: 1
  • Releases: 0
Created almost 6 years ago · Last pushed 10 months ago
Metadata Files
Readme License Citation

README.md

enpheeph

enpheeph - Neural Fault Injection Framework

We provide a customizable Fault Injection Framework, capable of injecting faults at different levels and with different libraries/networks.

NOTE: As of 2022-07-31, the framework is barely functional, hence there are still no official releases of the software. However, we plan on improving the framework over the course of the upcoming months to reach a point where it is stable enough to be used by researchers.

Installation

To install the currently work-in-progress version use:

pip install -e .

This will install the downloaded source in editable mode, hence it will be linked directly to the download location, rather than being a copy in the default site-packages folder in the Python distribution. It is highly recommended to use a virtual environment when installing enpheeph

NOTE: `enpheeph cannot be used without installing it, as the dependency checks and the imports are absolute, i.e. independent of where the actual folder is, but require the installation to be working.

NOTE: while enpheeph should automatically install all the requirements, a full list for the complete version can be found in papers/iros2022/requirements.txt

Development

Conventions

  1. We follow Conventional Commits for commit messages
  2. We use Semantic Versioning for versions

Tools

Install all the dependencies with

pip install -e .[dev]

Citation

enpheeph has been used for the following publications:

  1. enpheeph: A Fault Injection Framework for Spiking and Compressed Deep Neural Networks to appear at IROS 2022.
    1. A preprint version is available on arXiv, and it is also present in the repository citation button.

Owner

  • Name: Alessio Colucci
  • Login: Alexei95
  • Kind: user
  • Location: Vienna, Austria
  • Company: TU Wien

University Assistant, PhD student in TUWien. Hardware Designer and PC Enthusiast. @Alexei95AT on Telegram.

GitHub Events

Total
  • Delete event: 1
  • Push event: 65
  • Pull request event: 2
  • Create event: 1
Last Year
  • Delete event: 1
  • Push event: 65
  • Pull request event: 2
  • Create event: 1