https://zenodo.org/records/17764289

Skip to main
Zenodo home
[                    ]
Communities
My dashboard
Log in Sign up
Published November 29, 2025 | Version v1
Technical note Open

Entropica: 1024-mode unitary evolution with Born-rule readout,
trained on TinyStories in under 2 hours on a laptop GPU

Creators

  * Wallener, Damir

Description

Entropica is the first generative language model whose forward pass
is physically realizable as a passive linear-optical interferometer
(zero electrical power during inference).  
A 1024-mode, 32-layer unitary network using only Reck-scheme MZI
meshes and Born-rule readout learns coherent TinyStories-style
generation in under 1.8 hours on a single laptop GPU.  
Model is ready for a full optical implementation path demonstrated
with printed phase masks and a $30 laser pointer.  
All code, weights, and dataset generation scripts are public.

Files

Entropica - Zero-power optical language model.pdf

Files (326.4 kB)

                    Name                       Size    Download all
Entropica - Zero-power optical language       326.4   Preview 
model.pdf                                     kB      Download
md5:fb1c5584c2ff2fb41ff68d7abc766414

Additional details

Identifiers

URL
    https://github.com/dwallener/EntropicaPublic

Software

Repository URL  
    https://github.com/dwallener/EntropicaPublic
Programming language  
    Python
Development Status  
    Active

2K
Views
218
Downloads
Show more details

                  All versions This version
Views
                  2,252        2,252
Total views

Downloads
                  218          218
Total downloads

Data volume
                  78.0 MB      78.0 MB
Total data volume

More info on how stats are collected....

Versions

External resources

Indexed in
      + [ope]
        OpenAIRE

Communities

Keywords and subjects

Keywords

  * neuromorphic
  * Born rule
  * Reck scheme
  * linear optics
  * unitary neural network
  * zero-power inference
  * optical language model
  * photonic AI

Details

DOI
    10.5281/zenodo.17764289
    DOI Badge

    DOI

    10.5281/zenodo.17764289

    Markdown

    [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.17764289.svg)](https://doi.org/10.5281/zenodo.17764289)

    reStructuredText

    .. image:: https://zenodo.org/badge/DOI/10.5281/zenodo.17764289.svg
      :target: https://doi.org/10.5281/zenodo.17764289

    HTML

    <a href="https://doi.org/10.5281/zenodo.17764289"><img src="https://zenodo.org/badge/DOI/10.5281/zenodo.17764289.svg" alt="DOI"></a>

    Image URL

    https://zenodo.org/badge/DOI/10.5281/zenodo.17764289.svg

    Target URL

    https://doi.org/10.5281/zenodo.17764289

Resource type
    Technical note
Publisher
    Zenodo

Rights

License
    cc-by-4.0 icon Creative Commons Attribution 4.0 International
    The Creative Commons Attribution license allows re-distribution
    and re-use of a licensed work on the condition that the creator
    is appropriately credited. Read more

Citation

Export

Technical metadata

Created
    November 29, 2025
Modified
    November 29, 2025

Jump up

About

  * About
  * Policies
  * Infrastructure
  * Principles
  * Projects
  * Roadmap
  * Contact

Blog

  * Blog

Help

  * FAQ
  * Docs
  * Guides
  * Support

Developers

  * REST API
  * OAI-PMH

Contribute

  * GitHub
  * Donate

Funded by

  * [cern]
  * [openair]
  * [eu]

Powered by CERN Data Centre & InvenioRDM

  * Status
  * Privacy policy
  * Cookie policy
  * Terms of Use
  * Support

This site uses cookies. Find out more on how we use cookies

Accept all cookies Accept only essential cookies