MCDNN (MNIST)

Closed weights IDSIA,SUPSI 2.7M parameters February 2012

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
IDSIA,SUPSI
Organisation type
Academia,Academia
Country
Switzerland
Published
13 February 2012
Authors
Dan Cireşan, Ueli Meier, Juergen Schmidhuber

What it does

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

Domain
Vision
Task
Character recognition (OCR), Image classification

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
2.7M

"We train five DNN columns per normalization, resulting in a total of 35 columns for the entire MCDNN. [Each DNN has an architecture] 1x29x29-20C4-MP2-40C5-MP3-150N-10N DNN" Parameter calculation: 20*4*4+20*40*5*5+150*40*3*3+10*150=75820 35*75820=2653700

Training data
2,100,000 tokens

The MNIST database contains 60,000 training images and 10,000 testing images (Wikipedia)

Epochs
800

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
1.6 × 10¹⁶ FLOP

Num of multiply-adds per forward pass 2 FLOPs/mult-add 3 (fp+bp FLOPs / fp FLOPs) 800 epochs 60.000 training size 35 networks "Training a DNN takes almost 14 hours and after 500 training epochs little additional improvement is observed" 500*60000*3*174147400=1.57e+16

How it was established
Operation counting

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Closed — provider access only
Model access
Unreleased
Training code
Unreleased

How it is classified

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

Record confidence
Likely
Citations
4,828

Sources

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

Reference
Multi-column Deep Neural Networks for Image Classification
Last updated
11 February 2026

What the numbers mean

Background

MCDNN (MNIST) was published by IDSIA,SUPSI, in Switzerland, in February 2012. It comes out of academia,Academia.

It works in Vision, and is recorded as doing character recognition (OCR), Image classification.

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

How it was trained

Producing it required around 1.6 × 10¹⁶ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

Around 2,100,000 tokens went into training it.

Answers

MCDNN (MNIST) — common questions

01

Is MCDNN (MNIST) open source?

No. MCDNN (MNIST) has not had its weights published, so it exists only as a service controlled by its owner.

02

How many parameters does MCDNN (MNIST) have?

MCDNN (MNIST) has 2.7M parameters. "We train five DNN columns per normalization, resulting in a total of 35 columns for the entire MCDNN. [Each DNN has an architecture] 1x29x29-20C4-MP2-40C5-MP3-150N-10N DNN" Parameter calculation: 20*4*4+20*40*5*5+150*40*3*3+10*150=75820 35*75820=2653700. 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.

03

Who created MCDNN (MNIST)?

MCDNN (MNIST) was published by IDSIA,SUPSI, based in Switzerland, categorised as academia,Academia.

04

When was MCDNN (MNIST) released?

MCDNN (MNIST) was published in February 2012. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

05

What is MCDNN (MNIST) used for?

MCDNN (MNIST) works in Vision, and is recorded as handling character recognition (OCR), Image classification. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

06

How much compute was used to train MCDNN (MNIST)?

Around 1.6 × 10¹⁶ 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.

07

What GPU do I need to run MCDNN (MNIST)?

None. MCDNN (MNIST) 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.

Source

Original publication

Record last updated 11 February 2026

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.