Coder Social home page Coder Social logo

raulmurillo / low-precision-emacs Goto Github PK

View Code? Open in Web Editor NEW

This project forked from craymichael/low-precision-emacs

0.0 0.0 0.0 99 KB

RTL design of posit, fixed-point, and floating point low-precision exact-multiply-and-accumulate (EMAC) units for deep neural networks (DNNs).

License: GNU General Public License v3.0

VHDL 100.00%

low-precision-emacs's Introduction

Low-Precision-EMACs

RTL design of posit, fixed-point, and floating point low-precision exact-multiply-and-accumulate (EMAC) units for deep neural networks (DNNs). These designs compute the underlying operation in all DNNs while preserving as much accuracy locally as possible by incorporating the Kulisch exact-dot-product.

This work was performed while a part of the Neuromorphic Artificial Intelligence (Nu.AI) Lab in the Department of Computer Engineering, Rochester Institute of Technology. If you intend to use this source, please cite one the following papers below. Note that the source code provided here deviates from that used in the papers, but exhibits largely the same functionalities. Also, note the license file in this repository regardless of your potential application.

@inproceedings{carmichael2019positron,
  title         = {Deep Positron: {A} Deep Neural Network Using the Posit Number System},
  author        = {Zachariah Carmichael and Hamed Fatemi Langroudi and Char Khazanov and Jeffrey Lillie and John L. Gustafson and Dhireesha Kudithipudi},
  pages         = {1421--1426},
  doi           = {10.23919/DATE.2019.8715262},
  url           = {https://doi.org/10.23919/DATE.2019.8715262},
  crossref      = {DBLP:conf/date/2019}
}

@proceedings{DBLP:conf/date/2019,
  title         = dateconf,
  year          = 2019,
  month         = mar,
  booktitle     = dateconf,
  publisher     = ieee,
  address       = {Florence, Italy},
  isbn          = {978-3-9819263-2-3},
  url           = {http://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=8704855},
  eventdate     = {2019-03-25/2019-03-29}
}
@inproceedings{carmichael2019performance,
  title        = {Performance-Efficiency Trade-off of Low-Precision Numerical Formats in Deep Neural Networks},
  author       = {Carmichael, Zachariah and Langroudi, Hamed F. and Khazanov, Char and Lillie, Jeffrey and Gustafson, John L. and Kudithipudi, Dhireesha},
  year         = 2019,
  booktitle    = {Proceedings of the Conference for Next Generation Arithmetic 2019},
  location     = {Singapore, Singapore},
  publisher    = {ACM},
  address      = {New York, NY, USA},
  series       = {CoNGA'19},
  pages        = {3:1--3:9},
  doi          = {10.1145/3316279.3316282},
  isbn         = {978-1-4503-7139-1},
  url          = {http://doi.acm.org/10.1145/3316279.3316282},
  articleno    = 3,
  numpages     = 9,
  acmid        = 3316282,
}

low-precision-emacs's People

Contributors

craymichael avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.