Attend-Infer-Repeat
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
- Google DeepMind
- Organisation type
- Industry
- Country
- United States of America
- Published
- 12 August 2016
- Authors
- SM Ali Eslami, Nicolas Heess, Theophane Weber, Yuval Tassa, David Szepesvari, Geoffrey E Hinton
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Vision
- Task
- Object recognition
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
- 82.1M
- Training data
- tokens
from appendix E: The convolutional neural network uses a 64×(5×5)-64×(5×5)-64×(5×5)-512 architecture.
60000 MNIST images
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
- 6.4 × 10¹⁶ FLOP
- How it was established
- Hardware
(peak FLOPs for GPU - 1244 GFLOPs) times (training time - 3600 * 48 second) * (0.3 assumed utilization rate)
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 Quadro K4000
- Wall-clock time
- 48 hours
48 hours for MNIST model, 72 hours for 3D scenes model
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Confident
- Citations
- 581
Sources
Where this record came from and when it was last checked.
- Reference
- Attend, Infer, Repeat: Fast Scene Understanding with Generative Models
- Last updated
- 25 May 2026
What the numbers mean
What this model is
Attend-Infer-Repeat was published by Google DeepMind, in the country recorded as United States of America, during August 2016. The publishing organisation is categorised as industry.
It works in the domain of Vision, and is recorded as performing the task of object recognition.
This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.
Training and provenance
Training it took a computation budget of roughly 6.4 × 10¹⁶ FLOP, on hardware recorded as NVIDIA Quadro K4000. That figure measures what producing the model cost, and has no bearing on how fast it answers.
Answers
Attend-Infer-Repeat — common questions
Attend-Infer-Repeat— how many parameters does it have?
It has a parameter count of 82.1M. from appendix E: The convolutional neural network uses a 64×(5×5)-64×(5×5)-64×(5×5)-512 architecture. 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.
Attend-Infer-Repeat— who created it?
It was published by Google DeepMind, based in United States of America, an organisation categorised as industry.
Attend-Infer-Repeat— when was it released?
It was published in August 2016. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
Attend-Infer-Repeat— what is it used for?
It works in the domain of Vision, and is recorded as handling the task of object recognition. These are the areas it was designed around; they describe intent rather than a hard boundary.
Attend-Infer-Repeat— how much compute was used to train it?
Training consumed around 6.4 × 10¹⁶ FLOP, on hardware recorded as NVIDIA Quadro K4000. 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.
Attend-Infer-Repeat— what GPU do I need to run it?
None. This 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.
Attend-Infer-Repeat— is it open source?
The licensing was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.
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.