Mesh-TensorFlow Transformer 2.9B (translation)

Closed weights Google Brain 2.9B parameters November 2018

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
Google Brain
Organisation type
Industry
Country
United States of America
Published
5 November 2018
Authors
Noam Shazeer, Youlong Cheng, Niki Parmar, Dustin Tran, Ashish Vaswani, Penporn Koanantakool, Peter Hawkins, Hyoukjoong Mingsheng Lee, Cliff Hong, Ryan Young, Blake Sepassi, Hechtman

What it does

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

Domain
Language
Task
Language modeling/generation, Translation

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
2.9B

2.9B from section 9.1 : "On the WMT14 En-Fr translation tasks (3), we trained the models for 3 epochs. The largest model (2.9B parameters) was trained for 22 hours on a 128-core TPUv2 cluster."

Training data
1,550,000,000 tokens

Per Attention is All You Need, WMT 2014 En-Fr is ~36 million sentence pairs. If the average sentence is ~25 tokens (ballpark), dataset size is 36M * 25 * 2 = 1.8B tokens

Epochs
10

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
6.8 × 10¹⁹ FLOP

flops = (64) * ( 45 * 10**12) * (22 * 3600) * (0.3) = 6.8e19 (num gpu) * (peak flops) * (time in seconds) * (assumed utilization rate) from section 9.1 : "On the WMT14 En-Fr translation tasks (3), we trained the models for 3 epochs. The largest model (2.9B parameters) was trained for 22 hours on a 128-core TPUv2 cluster." from https://en.wikipedia.org/wiki/Tensor_Processing_Unit 45TFLOPs per chips

How it was established
Hardware

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
Google TPU v2
Chips used
64
Chip-hours
1,408
Wall-clock time
22 hours

from section 9.1 "On the WMT14 En-Fr translation tasks (3), we trained the models for 3 epochs. The largest model (2.9B parameters) was trained for 22 hours on a 128-core TPUv2 cluster."

Power draw
37.0 kW
Compute cost
$396

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

code here, apache license https://github.com/tensorflow/mesh/tree/master/mesh_tensorflow/transformer https://github.com/tensorflow/mesh/blob/master/mesh_tensorflow/transformer/transformer.py

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
SOTA improvement

'Using TPU meshes of up to 512 cores, we train Transformer models with up to 5 billion parameters, surpassing state of the art results on WMT'14 English-to-French translation task and the one-billion-word language modeling benchmark.'

Record confidence
Likely
Citations
437

Sources

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

Reference
Mesh-TensorFlow: Deep Learning for Supercomputers
Last updated
25 May 2026

What the numbers mean

Background

Mesh-TensorFlow Transformer 2.9B (translation) was published by Google Brain, in United States of America, in November 2018. industry is the category the publisher falls under.

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

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 6.8 × 10¹⁹ FLOP of arithmetic, on Google TPU v2, which is a statement about the training budget rather than about inference.

Around 1,550,000,000 tokens went into training it.

It is tracked in the underlying dataset for one reason in particular: sOTA improvement.

Answers

Mesh-TensorFlow Transformer 2.9B (translation) — common questions

01

How many parameters does Mesh-TensorFlow Transformer 2.9B (translation) have?

Mesh-TensorFlow Transformer 2.9B (translation) has 2.9B parameters. 2.9B from section 9.1 : "On the WMT14 En-Fr translation tasks (3), we trained the models for 3 epochs. The largest model (2.9B parameters) was trained for 22 hours on a 128-core TPUv2 cluster.". 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 Mesh-TensorFlow Transformer 2.9B (translation)?

Mesh-TensorFlow Transformer 2.9B (translation) was published by Google Brain, based in United States of America, categorised as industry.

03

When was Mesh-TensorFlow Transformer 2.9B (translation) released?

Mesh-TensorFlow Transformer 2.9B (translation) was published in November 2018. 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 Mesh-TensorFlow Transformer 2.9B (translation) used for?

Mesh-TensorFlow Transformer 2.9B (translation) works in Language, and is recorded as handling language modeling/generation, Translation. These are the areas it was designed around; they describe intent rather than a hard boundary.

05

How much compute was used to train Mesh-TensorFlow Transformer 2.9B (translation)?

Around 6.8 × 10¹⁹ FLOP, on Google TPU v2. 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 Mesh-TensorFlow Transformer 2.9B (translation)?

None. Mesh-TensorFlow Transformer 2.9B (translation) 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 Mesh-TensorFlow Transformer 2.9B (translation) open source?

No. Mesh-TensorFlow Transformer 2.9B (translation) has not had its weights published, so it exists only as a service controlled by its owner.

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.