Variational (untied weights, MC) LSTM (Large)

Closed weights University of Cambridge 66M parameters December 2015

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
University of Cambridge
Organisation type
Academia
Country
United Kingdom of Great Britain and Northern Ireland
Published
16 December 2015
Authors
Yarin Gal, Zoubin Ghahramani

What it does

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

Domain
Language
Task
Language modeling
Numerical format
FP32

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
66M

66M according to https://arxiv.org/pdf/1611.01462

Training data
929,000 tokens

"We had to use early stopping for the large model with [20]’s variant as the model starts overfitting after 16 epochs."

Epochs
16

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
5.9 × 10¹⁵ FLOP

6 FLOP / parameter / token * 66000000 parameters * 929000 tokens * 16 epochs = 5.886144e+15 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
Unreleased

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
Highly cited,SOTA improvement

"The new approach outperforms existing techniques, and to the best of our knowledge improves on the single model state-of-the-art in language modelling with the Penn Treebank (73.4 test perplexity)"

Record confidence
Confident
Citations
1,838
Benchmark data
Variational (untied weights, MC) LSTM (Large)

Sources

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

Reference
A Theoretically Grounded Application of Dropout in Recurrent Neural Networks
Last updated
11 February 2026

What the numbers mean

What this model is

Variational (untied weights, MC) LSTM (Large) was published by University of Cambridge, in United Kingdom of Great Britain and Northern Ireland, in December 2015. It comes out of 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 5.9 × 10¹⁵ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

It was trained on about 929,000 tokens of text.

Its inclusion criterion is highly cited,SOTA improvement.

Answers

Variational (untied weights, MC) LSTM (Large) — common questions

01

How much compute was used to train Variational (untied weights, MC) LSTM (Large)?

Around 5.9 × 10¹⁵ 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.

02

What GPU do I need to run Variational (untied weights, MC) LSTM (Large)?

None. Variational (untied weights, MC) LSTM (Large) 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.

03

Is Variational (untied weights, MC) LSTM (Large) open source?

No. Variational (untied weights, MC) LSTM (Large) has not had its weights published, so it exists only as a service controlled by its owner.

04

How many parameters does Variational (untied weights, MC) LSTM (Large) have?

Variational (untied weights, MC) LSTM (Large) has 66M parameters. 66M according to https://arxiv.org/pdf/1611.01462. 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.

05

Who created Variational (untied weights, MC) LSTM (Large)?

Variational (untied weights, MC) LSTM (Large) was published by University of Cambridge, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia.

06

When was Variational (untied weights, MC) LSTM (Large) released?

Variational (untied weights, MC) LSTM (Large) was published in December 2015. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

07

What is Variational (untied weights, MC) LSTM (Large) used for?

Variational (untied weights, MC) LSTM (Large) 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.

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.