Adaptive Inputs + LayerDrop TPS calculator

Open weights Facebook AI Research,LORIA 423M parameters September 2019

Each card below is assessed against this model at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from the card's memory bandwidth and the size of the model once compressed.

Calculated for this model

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · Q8_0 · 87.2 tok/s

Fastest card

B200

8,010 tok/s · 180 GB

Which GPUs can run Adaptive Inputs + LayerDrop?

Set the inputs, read the answer

A longer conversation needs more memory, which can push this model off smaller cards.

Hides cards that would only fit the model by compressing it below this point.

818 cards match

Calculating
Needs Quantisation Fit
8,010 tok/s

4,806–12,816 · low confidence

B200 NVIDIA 180 GB 8,000 GB/s Jan 2024 1.2 GB Q8_0 Comfortable
8,010 tok/s

4,806–12,816 · low confidence

B300 NVIDIA 288 GB 8,000 GB/s Sep 2025 1.2 GB Q8_0 Comfortable
6,396 tok/s

3,838–10,234 · low confidence

Radeon Instinct MI350X AMD 288 GB 8,190 GB/s Jan 2025 1.2 GB Q8_0 Comfortable
6,396 tok/s

3,838–10,234 · low confidence

Radeon Instinct MI355X AMD 288 GB 8,190 GB/s Jan 2025 1.2 GB Q8_0 Comfortable
5,115 tok/s

3,069–8,185 · low confidence

Radeon Instinct MI300 AMD 128 GB 6,550 GB/s Jan 2023 1.2 GB Q8_0 Comfortable
4,896 tok/s

2,938–7,834 · low confidence

H200 NVL NVIDIA 141 GB 4,890 GB/s Nov 2024 1.2 GB Q8_0 Comfortable
4,896 tok/s

2,938–7,834 · low confidence

H200 SXM 141 GB NVIDIA 141 GB 4,890 GB/s Nov 2024 1.2 GB Q8_0 Comfortable
4,686 tok/s

2,812–7,497 · low confidence

Radeon Instinct MI325X AMD 256 GB 6,000 GB/s Oct 2024 1.2 GB Q8_0 Comfortable
4,159 tok/s

2,495–6,654 · low confidence

Radeon Instinct MI300A AMD 128 GB 5,325 GB/s Dec 2023 1.2 GB Q8_0 Comfortable
4,159 tok/s

2,495–6,654 · low confidence

Radeon Instinct MI300X AMD 192 GB 5,325 GB/s Dec 2023 1.2 GB Q8_0 Comfortable
4,159 tok/s

2,495–6,654 · low confidence

Radeon Instinct MI308X AMD 192 GB 5,325 GB/s Dec 2023 1.2 GB Q8_0 Comfortable
3,945 tok/s

2,367–6,312 · low confidence

H100 NVL 94 GB NVIDIA 94 GB 3,940 GB/s Mar 2023 1.2 GB Q8_0 Comfortable
3,364 tok/s

2,019–5,383 · low confidence

H100 PCIe 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.2 GB Q8_0 Comfortable
3,364 tok/s

2,019–5,383 · low confidence

H100 SXM5 80 GB NVIDIA 80 GB 3,360 GB/s Oct 2022 1.2 GB Q8_0 Comfortable
3,364 tok/s

2,019–5,383 · low confidence

H100 SXM5 94 GB NVIDIA 94 GB 3,360 GB/s Mar 2023 1.2 GB Q8_0 Comfortable
3,364 tok/s

2,019–5,383 · low confidence

H100 SXM5 96 GB NVIDIA 96 GB 3,360 GB/s Mar 2023 1.2 GB Q8_0 Comfortable
3,364 tok/s

2,019–5,383 · low confidence

H800 SXM5 NVIDIA 80 GB 3,360 GB/s Mar 2023 1.2 GB Q8_0 Comfortable
2,562 tok/s

1,537–4,099 · low confidence

Radeon Instinct MI250 AMD 128 GB 3,280 GB/s Nov 2021 1.2 GB Q8_0 Comfortable
2,562 tok/s

1,537–4,099 · low confidence

Radeon Instinct MI250X AMD 128 GB 3,280 GB/s Nov 2021 1.2 GB Q8_0 Comfortable
2,135 tok/s

1,281–3,415 · low confidence

Data Center GPU Max 1550 Intel 128 GB 3,280 GB/s Jan 2023 1.2 GB Q8_0 Comfortable
2,089 tok/s

1,253–3,343 · low confidence

Data Center GPU Max Subsystem Intel 128 GB 3,210 GB/s Jan 2023 1.2 GB Q8_0 Comfortable
2,043 tok/s

1,226–3,268 · low confidence

A100 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Nov 2020 1.2 GB Q8_0 Comfortable
2,043 tok/s

1,226–3,268 · low confidence

A100X NVIDIA 80 GB 2,040 GB/s Jun 2021 1.2 GB Q8_0 Comfortable
2,043 tok/s

1,226–3,268 · low confidence

A800 SXM4 80 GB NVIDIA 80 GB 2,040 GB/s Aug 2022 1.2 GB Q8_0 Comfortable
2,043 tok/s

1,226–3,268 · low confidence

H100 CNX NVIDIA 80 GB 2,040 GB/s Mar 2023 1.2 GB Q8_0 Comfortable

Speeds are estimates for a single request — one conversation at a time — calculated from memory bandwidth, model size and quantisation. Real throughput varies with the inference runtime and its version. Figures published by hardware vendors measure many simultaneous requests and are much higher.

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,LORIA
Organisation type
Industry,Academia
Country
United States of America, France
Published
25 September 2019
Authors
Angela Fan, Edouard Grave, 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, Translation, Question answering, Text summarization, 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
423M
Training data
103,000,000 tokens

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
Open — downloadable
Model access
Open weights (unrestricted)
Training code
Open source

https://github.com/facebookresearch/fairseq/blob/main/examples/layerdrop/README.md Repo has MIT license WT training: https://github.com/facebookresearch/fairseq/tree/main/examples/language_model

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

"In neural machine translation on newstest2014, our 12 encoder layer Transformer model with LayerDrop further improves the state of the art, reaching 30.2 BLEU"

Record confidence
Confident
Citations
702
Benchmark data
Adaptive Inputs + LayerDrop

Sources

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

Reference
Reducing Transformer Depth on Demand with Structured Dropout
Last updated
25 May 2026

The extremes

What the numbers mean

The hardware side

Minimum card

Tesla C1080

Memory needed

1.2 GB

Fastest

8,010 tok/s

Adaptive Inputs + LayerDrop is small enough at 423M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 87.2 tokens per second.

Top of the range is the B200, at roughly 8,010 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

Adaptive Inputs + LayerDrop was published by Facebook AI Research,LORIA, in United States of America, in September 2019. It comes out of industry,Academia.

It works in Language, and is recorded as doing language modeling/generation, Translation, Question answering, Text summarization, Language modeling.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.

Understanding the speeds

Across every card that can run it, the middle of the range is about 224.9 tokens per second, and 817 of them clear the ten tokens per second that roughly matches reading speed.

Being dense, it reads all of itself per token, which is why the ordering by speed below follows the ordering by memory bandwidth so closely.

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

How it was trained

It was trained on about 103,000,000 tokens of text.

Its inclusion criterion is sOTA improvement.

Step by step

How to choose a GPU for Adaptive Inputs + LayerDrop

The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.

  1. 01

    Check what it needs before anything else

    Look at what Adaptive Inputs + LayerDrop actually needs — around 1.2 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.

  2. 02

    Set the context length you will work at

    Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason Adaptive Inputs + LayerDrop stops fitting a card that seemed fine.

  3. 03

    Set a quality floor

    Compression is what makes Adaptive Inputs + LayerDrop fit smaller cards, at some cost in accuracy — Q8_0 on the smallest card that fits. A minimum quality removes the ones that go too far.

  4. 04

    Sort by speed

    Sort by speed to see how cards rank for Adaptive Inputs + LayerDrop. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 8,010 tok/s.

  5. 05

    Check the fit verdict before buying

    Tight means Adaptive Inputs + LayerDrop loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.

  6. 06

    Open the card you have settled on

    Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for Adaptive Inputs + LayerDrop alone — a card is usually bought for more than one model.

Answers

Adaptive Inputs + LayerDrop — common questions

01

Can I run Adaptive Inputs + LayerDrop on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.2 GB and generating roughly 1,342 tokens per second — a comfortable fit.

02

Is Adaptive Inputs + LayerDrop open source?

Its weights are published, so Adaptive Inputs + LayerDrop can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.

03

How many parameters does Adaptive Inputs + LayerDrop have?

Adaptive Inputs + LayerDrop has 423M 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.

04

Who created Adaptive Inputs + LayerDrop?

Adaptive Inputs + LayerDrop was published by Facebook AI Research,LORIA, based in United States of America, categorised as industry,Academia.

05

When was Adaptive Inputs + LayerDrop released?

Adaptive Inputs + LayerDrop was published in September 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.

06

What is Adaptive Inputs + LayerDrop used for?

Adaptive Inputs + LayerDrop works in Language, and is recorded as handling language modeling/generation, Translation, Question answering, Text summarization, Language modeling. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

07

Where can I download Adaptive Inputs + LayerDrop?

The weights for Adaptive Inputs + LayerDrop are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

08

Can I run Adaptive Inputs + LayerDrop if it does not fit in my GPU?

Partly. Layers that do not fit sit in system memory and run at a fraction of the speed, so a mostly-offloaded Adaptive Inputs + LayerDrop is rarely worth using. Every figure here assumes the whole model is on the card.

09

Would two GPUs run Adaptive Inputs + LayerDrop faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run Adaptive Inputs + LayerDrop alone, the case for pairing is weak.

10

Why does the quantisation differ between cards for Adaptive Inputs + LayerDrop?

A larger card holds a more accurate copy. Across the cards that run Adaptive Inputs + LayerDrop, 1 compression levels are used; the floor control above pins it to one.

11

How accurate are these Adaptive Inputs + LayerDrop speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 4,806–12,816 tok/s on the B200, for instance. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

12

What GPU do I need to run Adaptive Inputs + LayerDrop?

The smallest card in our catalogue that holds Adaptive Inputs + LayerDrop is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.2 GB, and produces roughly 87.2 tokens per second. 818 cards in total can run it.

13

How fast is Adaptive Inputs + LayerDrop on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 8,010 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 817 of the cards that can run Adaptive Inputs + LayerDrop clear that.

14

How much VRAM does Adaptive Inputs + LayerDrop need?

About 1.2 GB at Q8_0 compression, which is what the smallest card that runs it uses. Less compression needs more: the figures in the memory column above are recalculated for each card, because each one holds the least-compressed version it can.

15

Can I run Adaptive Inputs + LayerDrop on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.2 GB and generating roughly 1,492 tokens per second — a comfortable fit.

16

Can I run Adaptive Inputs + LayerDrop on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.2 GB and generating roughly 914 tokens per second — a comfortable fit.

17

Can I run Adaptive Inputs + LayerDrop on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.2 GB and generating roughly 1,131 tokens per second — a comfortable fit.

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.