Handwritten digit recognition network

Closed weights AT&T 2.6K parameters November 1989

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
AT&T
Organisation type
Industry
Country
United States of America
Published
27 November 1989
Authors
Yann LeCun, B. Boser, J. Denker, D. Henderson, R. Howard, W. Hubbard, L. Jackel

What it does

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

Domain
Vision
Task
Digit recognition

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
2.6K

"In summary, the network has 4635 units, 98442 connections, and 2578 independent parameters.“

Training data
9,840 tokens

"After 30 training passes the error rate on training set (7291 handwritten plus 2549 printed digits)"

Epochs
30

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
1.8 × 10¹¹ FLOP

1.4e6 * 3 * 24 * 60* 60 * 0.5 = 181440000000 = 1.81e11 "A complete training session (30 passes through the training set plus test) takes about 3 days on a SUN SPARCstation 1" Sparcstation 1 has an estimated compute of 1.4 MFLOPS (source: https://ieeexplore.ieee.org/document/63671 )

How it was established
Hardware

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Wall-clock time
72 hours

"A complete training session (30 passes through the training set plus test) takes about 3 days on a SUN SPARCstation 1"

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
Handwritten Digit Recognition with a Back-Propagation Network
Last updated
28 November 2025

What the numbers mean

Where it came from

Handwritten digit recognition network was published by AT&T, in the country recorded as United States of America, during November 1989. The publishing organisation is categorised as industry.

It works in the domain of Vision, and is recorded as performing the task of digit recognition.

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

Training and provenance

The training run consumed about 1.8 × 10¹¹ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.

It was trained on a corpus of about 9,840 tokens of text.

The reason it appears in this catalogue at all: historical significance.

Answers

Handwritten digit recognition network — common questions

01

Handwritten digit recognition network— what GPU do I need to run it?

None. This 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.

02

Handwritten digit recognition network— is it open source?

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

03

Handwritten digit recognition network— how many parameters does it have?

It has a parameter count of 2.6K. "In summary, the network has 4635 units, 98442 connections, and 2578 independent parameters.“. 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.

04

Handwritten digit recognition network— who created it?

It was published by AT&T, based in United States of America, an organisation categorised as industry.

05

Handwritten digit recognition network— when was it released?

It was published in November 1989. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

06

Handwritten digit recognition network— what is it used for?

It works in the domain of Vision, and is recorded as handling the task of digit recognition. These are the areas it was designed around; they describe intent rather than a hard boundary.

07

Handwritten digit recognition network— how much compute was used to train it?

Training consumed around 1.8 × 10¹¹ 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.

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.