6-Layer-Tensor-Transformer+AdaHessian

Closed weights NERSC, Lawrence Berkeley National Laboratory,University of California (UC) Berkeley 85.5M parameters June 2020

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
NERSC, Lawrence Berkeley National Laboratory,University of California (UC) Berkeley
Organisation type
Government,Academia
Country
United States of America
Published
1 June 2020
Authors
Zhewei Yao, Amir Gholami, Sheng Shen, Mustafa Mustafa, Kurt Keutzer, Michael W. Mahoney

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
85.5M

" Following [35], a three-layer tensorized transformer core-1 for PTB and a six-layer tensorized transformer core-1 for Wikitext-103 are used in the experiments" [35] Ma, X.; Zhang, P.; Zhang, S.; Duan, N.; Hou, Y.; Zhou, M.; and Song, D. 2019. A tensorized transformer for language modeling. In Advances in Neural Information Processing Systems, 2229–2239. https://arxiv.org/pdf/1906.09777 Tensorized transformer has 85.5M parameters

Training data
103,000,000 tokens

batch size 120 10^7 steps (Figure 10)

Epochs
30

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
1.6 × 10¹⁸ FLOP

6 FLOP / parameter / token * 85500000 parameters * 103000000 tokens * 30 epochs = 1.58517e+18 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
Open source

MIT for code. doesn't have training script for WT103 but looks fairly adaptable: https://github.com/amirgholami/ADAHESSIAN

How it is classified

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

Record confidence
Confident
Citations
365
Benchmark data
6-Layer-Tensor-Transformer+AdaHessian

Sources

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

Reference
ADAHESSIAN: An Adaptive Second Order Optimizer for Machine Learning
Last updated
25 May 2026

What the numbers mean

Where it came from

6-Layer-Tensor-Transformer+AdaHessian was published by NERSC, Lawrence Berkeley National Laboratory,University of California (UC) Berkeley, in United States of America, in June 2020. It comes out of government,Academia.

It works in Language, and is recorded as doing language 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

Producing it required around 1.6 × 10¹⁸ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

The training set ran to roughly 103,000,000 tokens.

Answers

6-Layer-Tensor-Transformer+AdaHessian — common questions

01

Who created 6-Layer-Tensor-Transformer+AdaHessian?

6-Layer-Tensor-Transformer+AdaHessian was published by NERSC, Lawrence Berkeley National Laboratory,University of California (UC) Berkeley, based in United States of America, categorised as government,Academia.

02

When was 6-Layer-Tensor-Transformer+AdaHessian released?

6-Layer-Tensor-Transformer+AdaHessian was published in June 2020. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

03

What is 6-Layer-Tensor-Transformer+AdaHessian used for?

6-Layer-Tensor-Transformer+AdaHessian 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.

04

How much compute was used to train 6-Layer-Tensor-Transformer+AdaHessian?

Around 1.6 × 10¹⁸ 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.

05

What GPU do I need to run 6-Layer-Tensor-Transformer+AdaHessian?

None. 6-Layer-Tensor-Transformer+AdaHessian 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.

06

Is 6-Layer-Tensor-Transformer+AdaHessian open source?

No. 6-Layer-Tensor-Transformer+AdaHessian has not had its weights published, so it exists only as a service controlled by its owner.

07

How many parameters does 6-Layer-Tensor-Transformer+AdaHessian have?

6-Layer-Tensor-Transformer+AdaHessian has 85.5M parameters. " Following [35], a three-layer tensorized transformer core-1 for PTB and a six-layer tensorized transformer core-1 for Wikitext-103 are used in the experiments" [35] Ma, X.; Zhang, P.; Zhang, S.; Duan, N.; Hou, Y.; Zhou, M.; and Song, D. 2019. A tensorized transformer for language modeling. In Advances in Neural Information Processing Systems, 2229–2239. https://arxiv.org/pdf/1906.09777 Tensorized transformer has 85.5M parameters. 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.

Source

Original publication

Record last updated 25 May 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.