Image generation

Closed weights University of Amsterdam 784K parameters December 2013

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
University of Amsterdam
Organisation type
Academia
Country
Netherlands
Published
20 December 2013
Authors
DP Kingma, M Welling

What it does

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

Domain
Vision
Task
Image clustering

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
784K

"We trained generative models (decoders) and corresponding encoders (a.k.a. recognition models) having 500 hidden units in case of MNIST" 784*500*2=784000 (Ignoring latent dimension)

Training data
47,040,000 tokens

"We trained generative models of images from the MNIST and Frey Face datasets" MNIST has 60k images https://en.wikipedia.org/wiki/MNIST_database Frey Face has 2k images https://cs.nyu.edu/~roweis/data.html

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
4.8 × 10¹⁴ FLOP

From https://openai.com/blog/ai-and-compute/ Appendix "less than 0.0000055 pfs-days" (86400*10^15*0.0000055) Figure 2 shows evaluations with 10^8 training samples 6*784000*100000000=470400000000000

How it was established
Third-party estimation

How it is classified

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

Why it is tracked
Highly cited
Record confidence
Confident
Citations
21,760

Sources

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

Reference
Auto-Encoding Variational Bayes
Last updated
28 November 2025

What the numbers mean

Background

Image generation was published by University of Amsterdam, in Netherlands, in December 2013. It comes out of academia.

It works in Vision, and is recorded as doing image clustering.

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 4.8 × 10¹⁴ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

The training set ran to roughly 47,040,000 tokens.

It is tracked in the underlying dataset for one reason in particular: highly cited.

Answers

Image generation — common questions

01

How many parameters does Image generation have?

Image generation has 784K parameters. "We trained generative models (decoders) and corresponding encoders (a.k.a. recognition models) having 500 hidden units in case of MNIST" 784*500*2=784000 (Ignoring latent dimension). 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.

02

Who created Image generation?

Image generation was published by University of Amsterdam, based in Netherlands, categorised as academia.

03

When was Image generation released?

Image generation was published in December 2013. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

04

What is Image generation used for?

Image generation works in Vision, and is recorded as handling image clustering. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

05

How much compute was used to train Image generation?

Around 4.8 × 10¹⁴ 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.

06

What GPU do I need to run Image generation?

None. Image generation 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.

07

Is Image generation open source?

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

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.