Memformer (4 encoder + 16 decoder)

Closed weights UC Davis,Westlake University,Facebook AI 76.2M parameters October 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
UC Davis,Westlake University,Facebook AI
Organisation type
Academia,Academia,Industry
Country
United States of America, China
Published
14 October 2020
Authors
Qingyang Wu, Zhenzhong Lan, Kun Qian, Jing Gu, Alborz Geramifard, Zhou Yu

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
76.2M
Training data
103,000,000 tokens

batch size 128 steps: 150000 sequence length: 128-1024

Epochs
11.93

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
1.2 × 10¹⁹ FLOP

32400000000000 FLOP / second / GPU * 4 GPUs * 96 hours * 3600 sec / hour * 0.3 [assumed utilization] = 1.3436928e+19 FLOP time could be given for all language modeling not just for this model -> likely to be an upper bound

How it was established
Hardware

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware
NVIDIA Tesla V100 DGXS 16 GB,NVIDIA GeForce RTX 2080 Ti 11GB
Chips used
4
Wall-clock time
96 hours

"We trained our model on NVIDIA V100 16GB and 2080Ti 11GB. <..> The training for language modeling took approximately four days on four GPUs." 4 days = 96 hours

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
77
Benchmark data
Memformer (4 encoder + 16 decoder)

Sources

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

Reference
Memformer: A Memory-Augmented Transformer for Sequence Modeling
Last updated
25 May 2026

What the numbers mean

About this model

Memformer (4 encoder + 16 decoder) was published by UC Davis,Westlake University,Facebook AI, in United States of America, in October 2020. It comes out of academia,Academia,Industry.

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.

How it was trained

The training run consumed about 1.2 × 10¹⁹ FLOP, on NVIDIA Tesla V100 DGXS 16 GB,NVIDIA GeForce RTX 2080 Ti 11GB. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

The training set ran to roughly 103,000,000 tokens.

Answers

Memformer (4 encoder + 16 decoder) — common questions

01

Who created Memformer (4 encoder + 16 decoder)?

Memformer (4 encoder + 16 decoder) was published by UC Davis,Westlake University,Facebook AI, based in United States of America, categorised as academia,Academia,Industry.

02

When was Memformer (4 encoder + 16 decoder) released?

Memformer (4 encoder + 16 decoder) was published in October 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.

03

What is Memformer (4 encoder + 16 decoder) used for?

Memformer (4 encoder + 16 decoder) 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 Memformer (4 encoder + 16 decoder)?

Around 1.2 × 10¹⁹ FLOP, on NVIDIA Tesla V100 DGXS 16 GB,NVIDIA GeForce RTX 2080 Ti 11GB. 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 Memformer (4 encoder + 16 decoder)?

None. Memformer (4 encoder + 16 decoder) 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 Memformer (4 encoder + 16 decoder) open source?

No. Memformer (4 encoder + 16 decoder) has not had its weights published, so it exists only as a service controlled by its owner.

07

How many parameters does Memformer (4 encoder + 16 decoder) have?

Memformer (4 encoder + 16 decoder) has 76.2M 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.

Source

Original publication

Record last updated 25 May 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.