Multi-cell LSTM

Closed weights University of Hyderabad 7.2M parameters November 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
University of Hyderabad
Organisation type
Academia
Country
India
Published
15 November 2018
Authors
Thomas Cherian, Akshay Badola, Vineet Padmanabhan

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
7.2M

Based on the details in the paper, the number of parameters in the Multi-cell LSTM model can be calculated as follows: The model has 2 hidden LSTM layers, each with 1500 hidden units Each LSTM unit is a multi-cell LSTM with 10 memory cells per unit For a standard LSTM layer with n hidden units: W matrix: n x input_size U matrix: n x n 4 bias vectors of size n (for input, forget, cell, output gates) So for each multi-cell LSTM layer with 1500 units and 10 cells per unit: W matrix: 1500 x inp…

Training data
929,000 tokens

50 epochs (from Figure 4)

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 × 10¹⁵ FLOP

6 FLOP / parameter / token * 7200000 parameters * 929000 tokens * 50 epochs = 2.00664e+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
SOTA improvement

"The proposed multi-cell LSTM language models outperform the state-of-the-art results on well-known Penn Treebank (PTB) setup"

Record confidence
Likely
Citations
6
Benchmark data
Multi-cell LSTM

Sources

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

Reference
Multi-cell LSTM Based Neural Language Model
Last updated
28 November 2025

What the numbers mean

Where it came from

Multi-cell LSTM was published by University of Hyderabad, in India, in November 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

Training it took roughly 2 × 10¹⁵ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

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

The reason it appears in this catalogue at all is sOTA improvement.

Answers

Multi-cell LSTM — common questions

01

Who created Multi-cell LSTM?

Multi-cell LSTM was published by University of Hyderabad, based in India, categorised as academia.

02

When was Multi-cell LSTM released?

Multi-cell LSTM was published in November 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.

03

What is Multi-cell LSTM used for?

Multi-cell LSTM works in Language, and is recorded as handling language modeling. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

04

How much compute was used to train Multi-cell LSTM?

Around 2 × 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.

05

What GPU do I need to run Multi-cell LSTM?

None. Multi-cell LSTM 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.

06

Is Multi-cell LSTM open source?

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

07

How many parameters does Multi-cell LSTM have?

Multi-cell LSTM has 7.2M parameters. Based on the details in the paper, the number of parameters in the Multi-cell LSTM model can be calculated as follows: The model has 2 hidden LSTM layers, each with 1500 hidden units Each LSTM unit is a multi-cell LSTM with 10 memory cells per unit For a standard LSTM layer with n hidden units: W matrix: n x input_size U matrix: n x n 4 bias vectors of size n (for input, forget, cell, output gates) So for each multi-cell LSTM layer with 1500 units and 10 cells per unit: W matrix: 1500 x input_size U matrix: 1500 x 1500 4 bias vectors of size 1500 Number of parameters is same as standard LSTM layer For the 2 hidden layers: Input size for Layer 1: embedding dimension (estimated 300 in paper) Input size for Layer 2: 1500 (output of layer 1) Total params = Layer 1: 1500 x (300 + 1500 + 4) = 2,706,000 Layer 2: 1500 x (1500 + 1500 + 4) = 4,506,000 Total Parameters = 2,706,000 + 4,506,000 = 7,212,000 So the total number of parameters for the Multi-cell LSTM model with 2 layers of 1500 units and 10 cells per unit is approximately 7.2 million. 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.

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.