AFP+FPI (PTB)

Closed weights University of Sheffield 2M parameters June 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
University of Sheffield
Organisation type
Academia
Country
United Kingdom of Great Britain and Northern Ireland
Published
4 June 2021
Authors
Zhengxiong Wang, Anton Ragni

What it does

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

Domain
Language
Task
Language 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
2M

- RNN with 1 hidden layer - 200 hidden units - Vocabulary size = 10,000 The RNN layer will have: - Input weight matrix U: (vocab size x hidden size) = (10,000 x 200) = 2,000,000 parameters - Recurrent weight matrix W: (hidden size x hidden size) = (200 x 200) = 40,000 parameters - Bias vector b: (hidden size) = (200) = 200 parameters Total parameters: U = 2,000,000 W = 40,000 b = 200 Total parameters = U + W + b = 2,000,000 + 40,000 + 200 = 2,040,200 Therefore,…

Training data
912,344 tokens
Epochs
20

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
2.2 × 10¹⁴ FLOP

6 FLOP / parameter / token * 2,040,200 parameters * 912344 tokens * 20 epochs = 2.2336371e+14 FLOP

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
Benchmark data
AFP+FPI (PTB)

Sources

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

Reference
Approximate Fixed-Points in Recurrent Neural Networks
Last updated
11 February 2026

What the numbers mean

Background

AFP+FPI (PTB) was published by University of Sheffield, in United Kingdom of Great Britain and Northern Ireland, in June 2021. It comes out of academia.

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

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

How it was trained

The training run consumed about 2.2 × 10¹⁴ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Around 912,344 tokens went into training it.

Answers

AFP+FPI (PTB) — common questions

01

How many parameters does AFP+FPI (PTB) have?

AFP+FPI (PTB) has 2M parameters. - RNN with 1 hidden layer - 200 hidden units - Vocabulary size = 10,000 The RNN layer will have: - Input weight matrix U: (vocab size x hidden size) = (10,000 x 200) = 2,000,000 parameters - Recurrent weight matrix W: (hidden size x hidden size) = (200 x 200) = 40,000 parameters - Bias vector b: (hidden size) = (200) = 200 parameters Total parameters: U = 2,000,000 W = 40,000 b = 200 Total parameters = U + W + b = 2,000,000 + 40,000 + 200 = 2,040,200 Therefore, the total number of parameters with 200 hidden units and 10,000 vocab size is 2,040,200. _______ from the Algorithmic progress paper spreadsheet. 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 AFP+FPI (PTB)?

AFP+FPI (PTB) was published by University of Sheffield, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia.

03

When was AFP+FPI (PTB) released?

AFP+FPI (PTB) was published in June 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 AFP+FPI (PTB) used for?

AFP+FPI (PTB) works in Language, and is recorded as handling language modeling. 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 AFP+FPI (PTB)?

Around 2.2 × 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 AFP+FPI (PTB)?

None. AFP+FPI (PTB) 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 AFP+FPI (PTB) open source?

No. AFP+FPI (PTB) 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.