DEQ-Transformer (Post-LN) + Jacobian Regularisation TPS calculator

Open weights Carnegie Mellon University (CMU),Intel Labs 98M parameters June 2021

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 · 376 tok/s

Fastest card

B200

34,574 tok/s · 180 GB

Which GPUs can run DEQ-Transformer (Post-LN) + Jacobian Regularisation?

Set the inputs, read the answer

Context length tokens

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

Minimum quality

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

Search cards

Card memory

Also show cards that cannot run it

818 cards match

Calculating

					Needs	Quantisation	Fit
34,574 tok/s 20,744–55,318 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.8 GB	Q8_0	Comfortable
34,574 tok/s 20,744–55,318 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.8 GB	Q8_0	Comfortable
27,608 tok/s 16,565–44,173 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
27,608 tok/s 16,565–44,173 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.8 GB	Q8_0	Comfortable
22,080 tok/s 13,248–35,328 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
21,133 tok/s 12,680–33,813 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
21,133 tok/s 12,680–33,813 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.8 GB	Q8_0	Comfortable
20,226 tok/s 12,135–32,361 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.8 GB	Q8_0	Comfortable
17,950 tok/s 10,770–28,720 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
17,950 tok/s 10,770–28,720 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
17,950 tok/s 10,770–28,720 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.8 GB	Q8_0	Comfortable
17,028 tok/s 10,217–27,244 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
14,521 tok/s 8,713–23,234 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
14,521 tok/s 8,713–23,234 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.8 GB	Q8_0	Comfortable
14,521 tok/s 8,713–23,234 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
14,521 tok/s 8,713–23,234 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
14,521 tok/s 8,713–23,234 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.8 GB	Q8_0	Comfortable
11,057 tok/s 6,634–17,691 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
11,057 tok/s 6,634–17,691 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.8 GB	Q8_0	Comfortable
9,214 tok/s 5,528–14,742 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
9,017 tok/s 5,410–14,428 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.8 GB	Q8_0	Comfortable
8,816 tok/s 5,290–14,106 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.8 GB	Q8_0	Comfortable
8,816 tok/s 5,290–14,106 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.8 GB	Q8_0	Comfortable
8,816 tok/s 5,290–14,106 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.8 GB	Q8_0	Comfortable
8,816 tok/s 5,290–14,106 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.8 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: Carnegie Mellon University (CMU),Intel Labs
Organisation type: Academia,Industry
Country: United States of America
Published: 28 June 2021
Authors: Shaojie Bai, Vladlen Koltun, J. Zico Kolter

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: 98M
Training data: tokens
Epochs: 23

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: 2.9 × 10¹⁹ FLOP
How it was established: Operation counting,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 GeForce RTX 2080 Ti 11GB
Chips used: 4
Wall-clock time: 250 hours (10.4 days)
Power draw: 2.0 kW

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

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: 81
Benchmark data: DEQ-Transformer (Post-LN) + Jacobian Regularisation

Sources

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

Reference: Stabilizing Equilibrium Models by Jacobian Regularization
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run DEQ-Transformer (Post-LN) + Jacobian Regularisation

Ranked by estimated tokens per second, newest card first where speeds tie. Because generation is bound by memory bandwidth, this ordering follows bandwidth rather than any gaming benchmark.

The smallest GPUs that still run DEQ-Transformer (Post-LN) + Jacobian Regularisation

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

What the numbers mean

What it takes to run this model

Minimum card

Tesla C1080

Memory needed

0.8 GB

Fastest

34,574 tok/s

DEQ-Transformer (Post-LN) + Jacobian Regularisation is small enough at 98M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The smallest card that holds it is the Tesla C1080 with 4 GB, running it at Q8_0 and producing around 376 tokens per second.

Top of the range is the B200, at roughly 34,574 tokens per second thanks to 8,000 GB/s of bandwidth.

What this model is

DEQ-Transformer (Post-LN) + Jacobian Regularisation was published by Carnegie Mellon University (CMU),Intel Labs, in United States of America, in June 2021. It comes out of academia,Industry.

It works in Language, and is recorded as doing language modeling.

Published weights mean the model runs on your machine rather than someone else's, which is what makes the hardware question below answerable at all.

What decides the speed

The median result is around 970.8 tokens per second; 818 cards produce text faster than most people read it.

It is a dense model, so every parameter is read for every token produced. That makes speed track memory bandwidth almost exactly — a card with twice the bandwidth generates roughly twice as fast.

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

Producing it required around 2.9 × 10¹⁹ FLOP of arithmetic, on NVIDIA GeForce RTX 2080 Ti 11GB, which is a statement about the training budget rather than about inference.

Step by step

How to choose a GPU for DEQ-Transformer (Post-LN) + Jacobian Regularisation

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

01

Read the memory figure first

Look at what DEQ-Transformer (Post-LN) + Jacobian Regularisation actually needs — around 0.8 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason DEQ-Transformer (Post-LN) + Jacobian Regularisation stops fitting a card that seemed fine.
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy of DEQ-Transformer (Post-LN) + Jacobian Regularisation — Q8_0 on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Rank by throughput rather than spec sheet

Ranking by tokens per second for DEQ-Transformer (Post-LN) + Jacobian Regularisation follows memory bandwidth, not core counts, which is why the B200 tops it at 34,574 tok/s.
05

Look at the headroom, not just the fit

Tight means DEQ-Transformer (Post-LN) + Jacobian Regularisation 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.
06

Check the card from the other side

Following a card through to its own page shows every other model it can hold, which is the question that follows once DEQ-Transformer (Post-LN) + Jacobian Regularisation is settled.

Answers

DEQ-Transformer (Post-LN) + Jacobian Regularisation — common questions

Why does the quantisation differ between cards for DEQ-Transformer (Post-LN) + Jacobian Regularisation?

Because capacity varies, so does how hard DEQ-Transformer (Post-LN) + Jacobian Regularisation has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these DEQ-Transformer (Post-LN) + Jacobian Regularisation speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 20,744–55,318 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.

What GPU do I need to run DEQ-Transformer (Post-LN) + Jacobian Regularisation?

The smallest card in our catalogue that holds DEQ-Transformer (Post-LN) + Jacobian Regularisation is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.8 GB, and produces roughly 376 tokens per second. 818 cards in total can run it.

How fast is DEQ-Transformer (Post-LN) + Jacobian Regularisation on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 34,574 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 818 of the cards that can run DEQ-Transformer (Post-LN) + Jacobian Regularisation clear that.

How much VRAM does DEQ-Transformer (Post-LN) + Jacobian Regularisation need?

About 0.8 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.

Can I run DEQ-Transformer (Post-LN) + Jacobian Regularisation on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.8 GB and generating roughly 6,439 tokens per second — a comfortable fit.

Can I run DEQ-Transformer (Post-LN) + Jacobian Regularisation on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.8 GB and generating roughly 3,943 tokens per second — a comfortable fit.

Can I run DEQ-Transformer (Post-LN) + Jacobian Regularisation on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.8 GB and generating roughly 4,884 tokens per second — a comfortable fit.

Can I run DEQ-Transformer (Post-LN) + Jacobian Regularisation on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.8 GB and generating roughly 5,791 tokens per second — a comfortable fit.

Is DEQ-Transformer (Post-LN) + Jacobian Regularisation open source?

Its weights are published, so DEQ-Transformer (Post-LN) + Jacobian Regularisation 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.

How many parameters does DEQ-Transformer (Post-LN) + Jacobian Regularisation have?

DEQ-Transformer (Post-LN) + Jacobian Regularisation has 98M parameters. 98M (Table 1). 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 DEQ-Transformer (Post-LN) + Jacobian Regularisation?

DEQ-Transformer (Post-LN) + Jacobian Regularisation was published by Carnegie Mellon University (CMU),Intel Labs, based in United States of America, categorised as academia,Industry.

When was DEQ-Transformer (Post-LN) + Jacobian Regularisation released?

DEQ-Transformer (Post-LN) + Jacobian Regularisation was published in June 2021. 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 DEQ-Transformer (Post-LN) + Jacobian Regularisation used for?

DEQ-Transformer (Post-LN) + Jacobian Regularisation works in Language, and is recorded as handling language modeling. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download DEQ-Transformer (Post-LN) + Jacobian Regularisation?

The weights for DEQ-Transformer (Post-LN) + Jacobian Regularisation are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

How much compute was used to train DEQ-Transformer (Post-LN) + Jacobian Regularisation?

Around 2.9 × 10¹⁹ FLOP, on 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.

Can I run DEQ-Transformer (Post-LN) + Jacobian Regularisation if it does not fit in my GPU?

It can be split between the card and system memory, but DEQ-Transformer (Post-LN) + Jacobian Regularisation generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run DEQ-Transformer (Post-LN) + Jacobian Regularisation faster?

Two cards buy memory rather than speed. That matters for DEQ-Transformer (Post-LN) + Jacobian Regularisation only if one card cannot hold it — 818 can, so a second adds little.

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.

All AI models