LSTM with forget gates

Closed weights IDSIA 0.3K parameters January 1999

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
Organisation type
Academia
Country
Switzerland
Published
2 January 1999
Authors
F. A. Gers, J. Schmidhuber, and F. Cummins

What it does

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

Domain
Language
Task
Language Structure Modeling

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
0.3K

See Table 1

Training data
140,870,000 tokens

Training was stopped after at most 30000 training streams, each of which was ended when the first prediction error or the 100000th successive input symbol occurred NOTE this is a weird task. Not sure how to measure dataset size (#seqs? #symbols?)

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
Citations
6,283

Sources

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

Reference
Learning to forget: Continual prediction with LSTM
Last updated
28 November 2025

What the numbers mean

What this model is

LSTM with forget gates was published by IDSIA, in Switzerland, in January 1999. It comes out of academia.

It works in Language, and is recorded as doing language Structure Modeling.

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

Training and provenance

It was trained on about 140,870,000 tokens of text.

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

Answers

LSTM with forget gates — common questions

01

What is LSTM with forget gates used for?

LSTM with forget gates works in Language, and is recorded as handling language Structure Modeling. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

02

What GPU do I need to run LSTM with forget gates?

None. LSTM with forget gates 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.

03

Is LSTM with forget gates open source?

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

04

How many parameters does LSTM with forget gates have?

LSTM with forget gates has 0.3K parameters. See Table 1. 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.

05

Who created LSTM with forget gates?

LSTM with forget gates was published by IDSIA, based in Switzerland, categorised as academia.

06

When was LSTM with forget gates released?

LSTM with forget gates was published in January 1999. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

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.