LTE speaker verification system

Closed weights IBM 2.1K parameters November 1966

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
IBM
Organisation type
Industry
Country
United States of America
Published
1 November 1966
Authors
K. P. Li; J. E. Dammann; W. D. Chapman

What it does

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

Domain
Speech
Task
Speech recognition (ASR)

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
2.1K

2 connected systems, 1st level LTE and 2nd level LTE. 1st Level: 1810 parameters ("Thus, every 20 msec after the beginning of the utterance, the 15 filter amplitudes were each represented by a 12-bit code, resulting in a 180-bit time sample of the spectrum for that interval. Each time sample was fed to the first-level LTE's, which reduced it to a 10-bit code") 2nd Level: 251 parameters ("This resulted in a 250-bit input pattern to the second level for the first half-second of each utterance. Eac…

Training data
417 tokens

Split between both systems, 287 for 1st level, 130 for 2nd level.

Epochs
131

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
1.1 × 10⁸ FLOP

1st and 2nd level system are trained separately, multiple versions of both are trained, I chose the largest clearly described training runs. 1st level LTE compute: 2*1810*28700=103894000=1.04e8 1st level steps: 28700 ("Only 287 samples were selected to train the 10 LTE's. The same algorithm was used as that used with the 100-class gain. Two LTE's converged before 100 training passes.") 2nd level LTE compute: 2*251*4030=2023060=2e6 2nd level steps: 4030 (31 epochs, 130 training examples, see Ta…

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
Historical significance
Record confidence
Likely

Sources

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

Reference
Experimental Studies in Speaker Verification, Using an Adaptive System
Last updated
28 November 2025

What the numbers mean

Where it came from

LTE speaker verification system was published by IBM, in United States of America, in November 1966. The organisation is categorised as industry.

It works in Speech, and is recorded as doing speech recognition (ASR).

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

What went into building it

Producing it required around 1.1 × 10⁸ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

The training set ran to roughly 417 tokens.

The reason it appears in this catalogue at all is historical significance.

Answers

LTE speaker verification system — common questions

01

How many parameters does LTE speaker verification system have?

LTE speaker verification system has 2.1K parameters. 2 connected systems, 1st level LTE and 2nd level LTE. 1st Level: 1810 parameters ("Thus, every 20 msec after the beginning of the utterance, the 15 filter amplitudes were each represented by a 12-bit code, resulting in a 180-bit time sample of the spectrum for that interval. Each time sample was fed to the first-level LTE's, which reduced it to a 10-bit code") 2nd Level: 251 parameters ("This resulted in a 250-bit input pattern to the second level for the first half-second of each utterance. Each 250-bit pattern was then classified by the LTE into one of two classes"). 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 LTE speaker verification system?

LTE speaker verification system was published by IBM, based in United States of America, categorised as industry.

03

When was LTE speaker verification system released?

LTE speaker verification system was published in November 1966. 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 LTE speaker verification system used for?

LTE speaker verification system works in Speech, and is recorded as handling speech recognition (ASR). 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 LTE speaker verification system?

Around 1.1 × 10⁸ 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 LTE speaker verification system?

None. LTE speaker verification system 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 LTE speaker verification system open source?

The licensing for LTE speaker verification system 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.