Predictive Coding NN

Closed weights Technical University of Munich 206.9K parameters December 1994

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
Technical University of Munich
Organisation type
Academia
Country
Germany
Published
2 December 1994
Authors
J. Schmidhuber, Stefan Heil

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Language
Task
Language modeling

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
206.9K

5*80*430+430+430*80+80=206910 "P has nk input units and k output units. n is called the "time-window size" "Note that the time-window was quite small (n = 5)." "alphabet consisted of k = 80 possible characters" "P had 430 hidden units"

Training data
600,000 tokens

Training dataset: 15000*40=600000 "The training set for the predictor was given by a set of 40 articles from the newspaper Miinchner M erkur, each containing between 10000 and 20000 characters."

Epochs
25

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
1.9 × 10¹³ FLOP

2*206910*3*15000000=18621900000000=1.86e13 "The training phase consisted of 25 sweeps through the training set"

How it was established
Operation counting

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Frontier model
Yes
Why it is tracked
Historical significance
Record confidence
Confident

Sources

Where this record came from and when it was last checked.

Reference
Predictive Coding with Neural Nets: Application to Text Compression
Last updated
28 November 2025

What the numbers mean

Background

Predictive Coding NN was published by Technical University of Munich, in Germany, in December 1994. academia is the category the publisher falls under.

It works in Language, and is recorded as doing language modeling.

Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.

Training and provenance

Producing it required around 1.9 × 10¹³ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

Around 600,000 tokens went into training it.

Its inclusion criterion is historical significance.

Answers

Predictive Coding NN — common questions

01

When was Predictive Coding NN released?

Predictive Coding NN was published in December 1994. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

02

What is Predictive Coding NN used for?

Predictive Coding NN works in Language, and is recorded as handling language modeling. These are the areas it was designed around; they describe intent rather than a hard boundary.

03

How much compute was used to train Predictive Coding NN?

Around 1.9 × 10¹³ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

04

What GPU do I need to run Predictive Coding NN?

None. Predictive Coding NN is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

05

Is Predictive Coding NN open source?

The licensing for Predictive Coding NN was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

06

How many parameters does Predictive Coding NN have?

Predictive Coding NN has 206.9K parameters. 5*80*430+430+430*80+80=206910 "P has nk input units and k output units. n is called the "time-window size" "Note that the time-window was quite small (n = 5)." "alphabet consisted of k = 80 possible characters" "P had 430 hidden units". That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

07

Who created Predictive Coding NN?

Predictive Coding NN was published by Technical University of Munich, based in Germany, categorised as academia.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.