Variational RHN + WT (PTB)

Closed weights ETH Zurich,IDSIA 23M parameters July 2016

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
ETH Zurich,IDSIA
Organisation type
Academia,Academia
Country
Switzerland
Published
12 July 2016
Authors
Julian Georg Zilly, Rupesh Kumar Srivastava, Jan Koutník, Jürgen Schmidhuber

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
23M
Training data
tokens
Epochs
1,000

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
1.3 × 10¹⁷ FLOP

6*23000000*929000*1000=1.28202e+17

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: https://github.com/jzilly/RecurrentHighwayNetworks

How it is classified

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

Record confidence
Likely
Benchmark data
Variational RHN + WT

Sources

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

Reference
Recurrent Highway Networks
Last updated
11 February 2026

What the numbers mean

About this model

Variational RHN + WT (PTB) was published by ETH Zurich,IDSIA, in Switzerland, in July 2016. The organisation is categorised as academia,Academia.

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

Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.

What went into building it

The training run consumed about 1.3 × 10¹⁷ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Answers

Variational RHN + WT (PTB) — common questions

01

What GPU do I need to run Variational RHN + WT (PTB)?

None. Variational RHN + WT (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.

02

Is Variational RHN + WT (PTB) open source?

No. Variational RHN + WT (PTB) has not had its weights published, so it exists only as a service controlled by its owner.

03

How many parameters does Variational RHN + WT (PTB) have?

Variational RHN + WT (PTB) has 23M 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.

04

Who created Variational RHN + WT (PTB)?

Variational RHN + WT (PTB) was published by ETH Zurich,IDSIA, based in Switzerland, categorised as academia,Academia.

05

When was Variational RHN + WT (PTB) released?

Variational RHN + WT (PTB) was published in July 2016. 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

What is Variational RHN + WT (PTB) used for?

Variational RHN + WT (PTB) works in Language, and is recorded as handling language modeling. These are the areas it was designed around; they describe intent rather than a hard boundary.

07

How much compute was used to train Variational RHN + WT (PTB)?

Around 1.3 × 10¹⁷ 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.

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.