$\infty$-former (SM)

Closed weights Universidade de Lisboa (ULisboa),DeepMind 124M parameters September 2021

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
Universidade de Lisboa (ULisboa),DeepMind
Organisation type
Academia,Industry
Country
Portugal, United Kingdom of Great Britain and Northern Ireland
Published
1 September 2021
Authors
Pedro Henrique Martins, Zita Marinho, André F. T. Martins

What it does

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

Domain
Language
Task
Language modeling/generation
Base model
GPT-2 (124M)

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
124M

assuming same as the base model GPT-2 small

Training data
200,000,000 tokens
Epochs
1

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
1.2 × 10²² FLOP

7.936 × 10^20 FLOP [base model compute estimation, "Speculative" confidence] + 1.488e+17 FLOP [fine-tune compute] = 7.937488e+20 FLOP

How it was established
Operation counting
Fine-tuning compute
1.5 × 10¹⁷ FLOP

"To understand if long-term memories can be used to extend a pre-trained language model, we fine-tune GPT-2 small (Radford et al., 2019) on Wikitext103 (Merity et al., 2017) and a subset of PG-19 (Rae et al., 2019) containing the first 2,000 books (≈ 200 million tokens) of the training set" 6 FLOP / token / parameter * 124000000 parameters * 200000000 tokens * 1 epoch = 1.488e+17 FLOP

The training run

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

Training hardware
NVIDIA GeForce RTX 2080 Ti 11GB
Chips used
1
Power draw
277 W

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
Open (non-commercial)

fine-tune code (this is a GPT-2 finetune), no clear license: https://github.com/deep-spin/infinite-former

How it is classified

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

Record confidence
Speculative
Citations
31
Benchmark data
$\infty$-former (SM)

Sources

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

Reference
$\infty$-former: Infinite Memory Transformer
Last updated
11 February 2026

What the numbers mean

Background

$\infty$-former (SM) was published by Universidade de Lisboa (ULisboa),DeepMind, in Portugal, in September 2021. academia,Industry is the category the publisher falls under.

It works in Language, and is recorded as doing language modeling/generation.

Its starting point was GPT-2 (124M) — most models at this scale are adapted from an existing base rather than built from nothing.

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 1.2 × 10²² FLOP of arithmetic, on NVIDIA GeForce RTX 2080 Ti 11GB, which is a statement about the training budget rather than about inference.

Around 200,000,000 tokens went into training it.

Answers

$\infty$-former (SM) — common questions

01

How many parameters does $\infty$-former (SM) have?

$\infty$-former (SM) has 124M parameters. assuming same as the base model GPT-2 small. 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 $\infty$-former (SM)?

$\infty$-former (SM) was published by Universidade de Lisboa (ULisboa),DeepMind, based in Portugal, categorised as academia,Industry.

03

When was $\infty$-former (SM) released?

$\infty$-former (SM) was published in September 2021. 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 $\infty$-former (SM) used for?

$\infty$-former (SM) works in Language, and is recorded as handling language modeling/generation. 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.

05

How much compute was used to train $\infty$-former (SM)?

Around 1.2 × 10²² FLOP, on NVIDIA GeForce RTX 2080 Ti 11GB. 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 $\infty$-former (SM)?

None. $\infty$-former (SM) 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 $\infty$-former (SM) open source?

No. $\infty$-former (SM) has not had its weights published, so it exists only as a service controlled by its owner.

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.