DiffStk-MRNN

Closed weights Pennsylvania State University,Rochester Institute of Technology 1M parameters April 2020

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
Pennsylvania State University,Rochester Institute of Technology
Organisation type
Academia,Academia
Country
United States of America
Published
4 April 2020
Authors
Ankur Mali, Alexander Ororbia, Daniel Kifer, Clyde Lee Giles

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
1M

All models trained on this dataset consisted of 100 hidden units Okay, let's break this down: - RNN with 1 hidden layer - 100 hidden units - Vocabulary size = 10,000 The RNN layer will have: - Input weight matrix U: dimensions (vocab size x hidden size) = (10,000 x 100) = 1,000,000 parameters - Recurrent weight matrix W: dimensions (hidden size x hidden size) = (100 x 100) = 10,000 parameters - Bias vector b: (hidden size) = (100) = 100 parameters So the total parameters is: U = 1,000,0…

Training data
912,344 tokens
Epochs
50

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
2.8 × 10¹⁴ FLOP

6 FLOP / token / parameter * 1010000 parameters * 912 344 tokens * 50 epochs = 2.7644023e+14 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.

Record confidence
Likely
Citations
14
Benchmark data
DiffStk-MRNN

Sources

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

Reference
Recognizing Long Grammatical Sequences Using Recurrent Networks Augmented With An External Differentiable Stack
Last updated
1 December 2025

What the numbers mean

Background

DiffStk-MRNN was published by Pennsylvania State University,Rochester Institute of Technology, in United States of America, in April 2020. academia,Academia is the category the publisher falls under.

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

Training it took roughly 2.8 × 10¹⁴ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

Around 912,344 tokens went into training it.

Answers

DiffStk-MRNN — common questions

01

What is DiffStk-MRNN used for?

DiffStk-MRNN 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.

02

How much compute was used to train DiffStk-MRNN?

Around 2.8 × 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.

03

What GPU do I need to run DiffStk-MRNN?

None. DiffStk-MRNN 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.

04

Is DiffStk-MRNN open source?

No. DiffStk-MRNN has not had its weights published, so it exists only as a service controlled by its owner.

05

How many parameters does DiffStk-MRNN have?

DiffStk-MRNN has 1M parameters. All models trained on this dataset consisted of 100 hidden units Okay, let's break this down: - RNN with 1 hidden layer - 100 hidden units - Vocabulary size = 10,000 The RNN layer will have: - Input weight matrix U: dimensions (vocab size x hidden size) = (10,000 x 100) = 1,000,000 parameters - Recurrent weight matrix W: dimensions (hidden size x hidden size) = (100 x 100) = 10,000 parameters - Bias vector b: (hidden size) = (100) = 100 parameters So the total parameters is: U = 1,000,000 W = 10,000 b = 100 Total parameters = U + W + b = 1,000,000 + 10,000 + 100 = 1,010,100 Therefore, the total number of parameters for this 1-layer RNN with 100 hidden units and 10,000 vocab size is 1,010,100. 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.

06

Who created DiffStk-MRNN?

DiffStk-MRNN was published by Pennsylvania State University,Rochester Institute of Technology, based in United States of America, categorised as academia,Academia.

07

When was DiffStk-MRNN released?

DiffStk-MRNN was published in April 2020. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

Source

Original publication

Record last updated 1 December 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.