NPLM (AP News)

Closed weights University of Montreal / Université de Montréal 11.9M parameters March 2003

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 Montreal / Université de Montréal
Organisation type
Academia
Country
Canada
Published
15 March 2003
Authors
Yoshua Bengio, Réjean Ducharme, Pascal Vincent, Christian Jauvin

What it does

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

Domain
Language
Task
Text autocompletion

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
11.9M

"The number of free parameters is |V|(1 + nm + h) + h(1 + (n − 1)m) [...] For example, consider the following architecture used in the experiments on the AP (Associated Press) news data: the vocabulary size is |V| = 17,964, the number of hidden units is h = 60, the order of the model is n = 6, the number of word features is m = 100" AP News: n=6, h=60, m=100, |V|=17964 17964*(1+6*100+60)+60*(1+(6-1)*100)=11904264

Training data
13,994,528 tokens

"Comparative experiments were performed on the Brown corpus which is a stream of 1,181,041 words, from a large variety of English texts and books. The first 800,000 words were used for training, the following 200,000 for validation (model selection, weight decay, early stopping) and the remaining 181,041 for testing. The number of different words is 47,578 (including punctuation, distinguishing between upper and lower case, and including the syntactical marks used to separate texts and paragraph…

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
1.7 × 10¹⁵ FLOP

"For example, consider the following architecture used in the experiments on the AP (Associated Press) news data: the vocabulary size is |V| = 17,964, the number of hidden units is h = 60, the order of the model is n = 6, the number of word features is m = 100. The total number of numerical operations to process a single training example is approximately |V|(1+nm+h)+h(1+nm)+nm" AP News: n=6, h=60, m=100, |V|=17964, dataset=13994528, epochs=5 Forward FLOP: 17964*(1+6*100+60)+60*(1+6*100)+6*100=1…

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
Record confidence
Confident
Citations
7,637

Sources

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

Reference
A Neural Probabilistic Language Model
Last updated
11 February 2026

What the numbers mean

About this model

NPLM (AP News) was published by University of Montreal / Université de Montréal, in Canada, in March 2003. academia is the category the publisher falls under.

It works in Language, and is recorded as doing text autocompletion.

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

What went into building it

Producing it required around 1.7 × 10¹⁵ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

The training set ran to roughly 13,994,528 tokens.

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

Answers

NPLM (AP News) — common questions

01

When was NPLM (AP News) released?

NPLM (AP News) was published in March 2003. 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 NPLM (AP News) used for?

NPLM (AP News) works in Language, and is recorded as handling text autocompletion. These are the areas it was designed around; they describe intent rather than a hard boundary.

03

How much compute was used to train NPLM (AP News)?

Around 1.7 × 10¹⁵ 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 NPLM (AP News)?

None. NPLM (AP News) 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 NPLM (AP News) open source?

The licensing for NPLM (AP News) 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 NPLM (AP News) have?

NPLM (AP News) has 11.9M parameters. "The number of free parameters is |V|(1 + nm + h) + h(1 + (n − 1)m) [...] For example, consider the following architecture used in the experiments on the AP (Associated Press) news data: the vocabulary size is |V| = 17,964, the number of hidden units is h = 60, the order of the model is n = 6, the number of word features is m = 100" AP News: n=6, h=60, m=100, |V|=17964 17964*(1+6*100+60)+60*(1+(6-1)*100)=11904264. 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 NPLM (AP News)?

NPLM (AP News) was published by University of Montreal / Université de Montréal, based in Canada, categorised as academia.

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.