All-attention network + adaptive span
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
- Facebook AI Research
- Organisation type
- Industry
- Country
- United States of America, France
- Published
- 2 July 2019
- Authors
- Sainbayar Sukhbaatar, Edouard Grave, Guillaume Lample, Herve Jegou, Armand Joulin
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling/generation
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
- 133M
- Training data
- 103,000,000 tokens
- Epochs
- 31.8
- Batch size
- 16,384
133M (Table 3) For word level language modeling, we use a model with d = 512 and 36 layers, each with 8 heads and 2048 persistent vectors.
8k warmup steps A batch consists of 64 samples, each with 256 tokens Following Baevski and Auli [2] on WikiText-103, we use tied adaptive softmax and adaptive input with 3 clusters of size 20k, 40k and 200k assuming 200k steps: 200000*64*256/103000000 = 31.8 epochs
64*256
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
- How it was established
- Hardware,Operation counting
31330000000000 FLOP / sec / GPU [fp16] * 64 GPUs * 24 hours * 3600 sec/hour * 0.3 [assumed utilization] = 5.197271e+19 FLOP 6 FLOP / token / parameter * 200000 steps * 64 sequences per batch * 256 tokens per sequence * 133000000 parameters = 2.6148864e+18 FLOP sqrt(5.197271e+19*2.6148864e+18) = 1.1657733e+19 FLOP likely confidence: i am uncertain about number of steps and v100 precision _________ The Algorithmic Progress paper estimation was 4.6 × 10^19 FLOP
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 V100
- Chips used
- 64
- Wall-clock time
- 24 hours
- Power draw
- 39.5 kW
"Training large models takes about a day on 64 V100 GPUs."
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
- 156
- Benchmark data
- All-attention network + adaptive span
Sources
Where this record came from and when it was last checked.
- Reference
- Augmenting Self-attention with Persistent Memory
- Last updated
- 25 May 2026
What the numbers mean
About this model
All-attention network + adaptive span was published by Facebook AI Research, in United States of America, in July 2019. It comes out of industry.
It works in Language, and is recorded as doing language modeling/generation.
This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.
What went into building it
The training run consumed about 1.2 × 10¹⁹ FLOP, on NVIDIA V100. That figure describes the cost of creating it and has no bearing on how quickly it generates text.
It was trained on about 103,000,000 tokens of text.
Answers
All-attention network + adaptive span — common questions
How many parameters does All-attention network + adaptive span have?
All-attention network + adaptive span has 133M parameters. 133M (Table 3) For word level language modeling, we use a model with d = 512 and 36 layers, each with 8 heads and 2048 persistent vectors. 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.
Who created All-attention network + adaptive span?
All-attention network + adaptive span was published by Facebook AI Research, based in United States of America, categorised as industry.
When was All-attention network + adaptive span released?
All-attention network + adaptive span was published in July 2019. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is All-attention network + adaptive span used for?
All-attention network + adaptive span works in Language, and is recorded as handling language modeling/generation. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
How much compute was used to train All-attention network + adaptive span?
Around 1.2 × 10¹⁹ FLOP, on NVIDIA V100. 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.
What GPU do I need to run All-attention network + adaptive span?
None. All-attention network + adaptive span 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.
Is All-attention network + adaptive span open source?
No. All-attention network + adaptive span has not had its weights published, so it exists only as a service controlled by its owner.
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.