RNNLM + Dynamic KL Regularization

Closed weights Northwestern University 13.3M parameters January 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
Northwestern University
Organisation type
Academia
Country
United States of America
Published
1 January 2018
Authors
Thanapon Noraset, David Demeter, Doug Downey

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
13.3M

2-layer LSTM, 650 hidden units per layer, embedding size = 650. Penn Treebank standard vocab = ~10k words (from Zaremba et al. 2014) 2 * (4 * (650*650 + 650*650 + 650)) + (650*10000)+10000 = 13275200

Training data
tokens

Penn Treebank known to have 930k tokens

Epochs
20

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 × 10¹⁴ FLOP

((8*650*(650+650))*2)*2 * 930000 tokens/epoch * 20 epochs = 5.029e14. batch size assumed from prior literature

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

code, but not experiment code: https://github.com/northanapon/seqmodel/tree/aaai18

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
9
Benchmark data
RNNLM + Dynamic KL Regularization

Sources

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

Reference
Controlling Global Statistics in Recurrent Neural Network Text Generation
Last updated
28 November 2025

What the numbers mean

What this model is

RNNLM + Dynamic KL Regularization was published by Northwestern University, in United States of America, in January 2018. The organisation is categorised as 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.

Training and provenance

The training run consumed about 5 × 10¹⁴ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Answers

RNNLM + Dynamic KL Regularization — common questions

01

How many parameters does RNNLM + Dynamic KL Regularization have?

RNNLM + Dynamic KL Regularization has 13.3M parameters. 2-layer LSTM, 650 hidden units per layer, embedding size = 650. Penn Treebank standard vocab = ~10k words (from Zaremba et al. 2014) 2 * (4 * (650*650 + 650*650 + 650)) + (650*10000)+10000 = 13275200. 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 RNNLM + Dynamic KL Regularization?

RNNLM + Dynamic KL Regularization was published by Northwestern University, based in United States of America, categorised as academia.

03

When was RNNLM + Dynamic KL Regularization released?

RNNLM + Dynamic KL Regularization was published in January 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 RNNLM + Dynamic KL Regularization used for?

RNNLM + Dynamic KL Regularization 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.

05

How much compute was used to train RNNLM + Dynamic KL Regularization?

Around 5 × 10¹⁴ 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.

06

What GPU do I need to run RNNLM + Dynamic KL Regularization?

None. RNNLM + Dynamic KL Regularization 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 RNNLM + Dynamic KL Regularization open source?

No. RNNLM + Dynamic KL Regularization has not had its weights published, so it exists only as a service controlled by its owner.

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.