Weight Decay

Closed weights 8.4K parameters December 1991

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.

Published
2 December 1991
Authors
A. Krogh, J. Hertz

What it does

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

Domain
Speech
Task
Speech synthesis

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
8.4K

7*26*40+40+40*26+26=8386 "The network had 7 x 26 input units, 40 hidden units and 26 output units"

Training data
25,000 tokens

"It was trained on 400 to 5000 random words from the data base of around 20.000 words,"

Epochs
300

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
7.5 × 10¹⁰ FLOP

2*8386*3*1500000=75474000000=7.55e10 "It was trained on 400 to 5000 random words from the data base of around 20.000 words," "The top full line corresponds to the generalization error after 300 epochs"

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
Highly cited,Historical significance
Record confidence
Confident

Sources

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

Reference
A Simple Weight Decay Can Improve Generalization
Last updated
28 November 2025

What the numbers mean

Background

Weight Decay was published by its authors, in December 1991.

It works in Speech, and is recorded as doing speech synthesis.

This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.

What went into building it

Producing it required around 7.5 × 10¹⁰ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

It was trained on about 25,000 tokens of text.

The reason it appears in this catalogue at all is highly cited,Historical significance.

Answers

Weight Decay — common questions

01

When was Weight Decay released?

Weight Decay was published in December 1991. 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 Weight Decay used for?

Weight Decay works in Speech, and is recorded as handling speech synthesis. These are the areas it was designed around; they describe intent rather than a hard boundary.

03

How much compute was used to train Weight Decay?

Around 7.5 × 10¹⁰ 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 Weight Decay?

None. Weight Decay 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 Weight Decay open source?

The licensing for Weight Decay 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 Weight Decay have?

Weight Decay has 8.4K parameters. 7*26*40+40+40*26+26=8386 "The network had 7 x 26 input units, 40 hidden units and 26 output 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.

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.