RNMT+

Closed weights Google AI 378.9M parameters April 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 AI
Organisation type
Industry
Country
United States of America
Published
26 April 2018
Authors
Mia Xu Chen, Orhan Firat, Ankur Bapna, Melvin Johnson, Wolfgang Macherey, George Foster, Llion Jones, Niki Parmar, Mike Schuster, Zhifeng Chen, Yonghui Wu, Macduff Hughes

What it does

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

Domain
Language
Task
Translation
Approach
Supervised

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
378.9M

from Table 3 RNMT+

Training data
tokens

"We train our models on the standard WMT’14 En→Fr and En→De datasets that comprise 36.3M and 4.5M sentence pairs, respectively" "For RNMT+, we use sentence-level cross entropy loss."

Epochs
8.5

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
1.8 × 10¹⁹ FLOP

32 * 9.3 TFLOPS * (120 * 3600) * 0.3 = 3.86e19 (number of gpus) * (peak flops) * (seconds) * (assumed utilization rate) "All models were trained with synchronous training. RNMT+ and ConvS2S were trained with 32 NVIDIA P100 GPUs", training time is 120h from Table 1. Alternatively, typical inference is 2.81e10 FLOP, there are 36.3M examples in the dataset, and they train for 8.5 epochs: 2.81e10 * 36.3M * 8.5 = 8.7e18 Geometric mean: sqrt(3.86e19 * 8.7e18) = 1.83e19

How it was established
Hardware,Operation counting

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 P100
Chips used
32
Chip-hours
3,840
Wall-clock time
120 hours

from Table 1

Power draw
16.6 kW

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
504

Sources

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

Reference
The Best of Both Worlds: Combining Recent Advances in Neural Machine Translation
Last updated
28 November 2025

What the numbers mean

About this model

RNMT+ was published by Google AI, in the country recorded as United States of America, during April 2018. It comes out of an organisation categorised as industry.

It works in the domain of Language, and is recorded as performing the task of translation.

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

How it was trained

Training it took a computation budget of roughly 1.8 × 10¹⁹ FLOP, on hardware recorded as NVIDIA P100. That figure measures what producing the model cost, and has no bearing on how fast it answers.

Answers

RNMT+ — common questions

01

RNMT+— what GPU do I need to run it?

None. This 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.

02

RNMT+— is it open source?

No. Its weights have not been published, so it exists only as a service controlled by its owner.

03

RNMT+— how many parameters does it have?

It has a parameter count of 378.9M. from Table 3 RNMT+. 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.

04

RNMT+— who created it?

It was published by Google AI, based in United States of America, an organisation categorised as industry.

05

RNMT+— when was it released?

It was published in April 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.

06

RNMT+— what is it used for?

It works in the domain of Language, and is recorded as handling the task of translation. These are the areas it was designed around; they describe intent rather than a hard boundary.

07

RNMT+— how much compute was used to train it?

Training consumed around 1.8 × 10¹⁹ FLOP, on hardware recorded as NVIDIA P100. 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.

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.