Transformer (Adaptive Input Embeddings) WT103 TPS calculator

Open weights Facebook AI Research 247M parameters September 2018

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 · 149 tok/s

Fastest card

B200

13,718 tok/s · 180 GB

Which GPUs can run Transformer (Adaptive Input Embeddings) WT103?

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
13,718 tok/s 8,231–21,948 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.0 GB	Q8_0	Comfortable
13,718 tok/s 8,231–21,948 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.0 GB	Q8_0	Comfortable
10,954 tok/s 6,572–17,526 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.0 GB	Q8_0	Comfortable
10,954 tok/s 6,572–17,526 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.0 GB	Q8_0	Comfortable
8,760 tok/s 5,256–14,017 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.0 GB	Q8_0	Comfortable
8,385 tok/s 5,031–13,416 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.0 GB	Q8_0	Comfortable
8,385 tok/s 5,031–13,416 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.0 GB	Q8_0	Comfortable
8,025 tok/s 4,815–12,840 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.0 GB	Q8_0	Comfortable
7,122 tok/s 4,273–11,395 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.0 GB	Q8_0	Comfortable
7,122 tok/s 4,273–11,395 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.0 GB	Q8_0	Comfortable
7,122 tok/s 4,273–11,395 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.0 GB	Q8_0	Comfortable
6,756 tok/s 4,054–10,809 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.0 GB	Q8_0	Comfortable
5,761 tok/s 3,457–9,218 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.0 GB	Q8_0	Comfortable
5,761 tok/s 3,457–9,218 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.0 GB	Q8_0	Comfortable
5,761 tok/s 3,457–9,218 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.0 GB	Q8_0	Comfortable
5,761 tok/s 3,457–9,218 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.0 GB	Q8_0	Comfortable
5,761 tok/s 3,457–9,218 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.0 GB	Q8_0	Comfortable
4,387 tok/s 2,632–7,019 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.0 GB	Q8_0	Comfortable
4,387 tok/s 2,632–7,019 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.0 GB	Q8_0	Comfortable
3,656 tok/s 2,193–5,849 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.0 GB	Q8_0	Comfortable
3,578 tok/s 2,147–5,724 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.0 GB	Q8_0	Comfortable
3,498 tok/s 2,099–5,597 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.0 GB	Q8_0	Comfortable
3,498 tok/s 2,099–5,597 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.0 GB	Q8_0	Comfortable
3,498 tok/s 2,099–5,597 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.0 GB	Q8_0	Comfortable
3,498 tok/s 2,099–5,597 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.0 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
Organisation type: Industry
Country: United States of America, France
Published: 28 September 2018
Authors: Alexei Baevski, Michael Auli

What it does

The problem areas the model was built for. A model can carry several of each.

Domain: Language
Task: Language modeling
Numerical format: FP16

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: 247M
Training data: 100,000,000 tokens
Epochs: 180

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: 4.5 × 10¹⁹ FLOP
How it was established: Hardware,Operation counting

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: 8
Chip-hours: 4,288
Wall-clock time: 67 hours
Power draw: 5.0 kW
Compute cost: $2,881

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.

Why it is tracked: SOTA improvement
Record confidence: Confident
Citations: 433
Benchmark data: Transformer (Adaptive Input Embeddings)

Sources

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

Reference: Adaptive Input Representations for Neural Language Modeling
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run Transformer (Adaptive Input Embeddings) WT103

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 Transformer (Adaptive Input Embeddings) WT103

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

1.0 GB

Fastest

13,718 tok/s

Transformer (Adaptive Input Embeddings) WT103 reaches a parameter count of 247M. That is small enough that hardware is rarely the obstacle, including on cards several years old. The number of cards we track that can run it: 818.

The smallest card that holds it is Tesla C1080, with a memory capacity of 4 GB, running it at a compression of Q8_0 and producing around 149 tokens per second.

Top of the range is B200, generating roughly 13,718 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

Background

Transformer (Adaptive Input Embeddings) WT103 was published by Facebook AI Research, in the country recorded as United States of America, during September 2018. The category the publisher falls under is industry.

It works in the domain of Language, and is recorded as performing the task of language modeling.

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

Reading the throughput figures

Half the cards that hold it manage more than 385.2 tokens per second. Exceeding reading speed outright: 818 of them.

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

Training it took a computation budget of roughly 4.5 × 10¹⁹ FLOP, on hardware recorded as NVIDIA V100. That figure measures what producing the model cost, and has no bearing on how fast it answers.

It was trained on a corpus of about 100,000,000 tokens of text.

Its inclusion criterion: sOTA improvement.

Step by step

How to choose a GPU for Transformer (Adaptive Input Embeddings) WT103

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

01

Start from the memory column

Every card here has been checked against Transformer (Adaptive Input Embeddings) WT103, needing around 1.0 GB at a compression of Q8_0. That figure, not the headline performance of a card, is what decides whether it runs.
02

Match the context to your actual use

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for Transformer (Adaptive Input Embeddings) WT103.
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy, reaching a compression of Q8_0 on the smallest card that fits. Setting a minimum quality drops the cards that only manage it by squeezing further than you would want, and holds the comparison at one level.
04

Rank by throughput rather than spec sheet

Sort by speed to see how cards rank for Transformer (Adaptive Input Embeddings) WT103. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 13,718 tok/s.
05

Look at the headroom, not just the fit

The fit column separates cards that just manage it from those with room to spare, in the case of Transformer (Adaptive Input Embeddings) WT103. Comfortable means you can grow the context later. That difference matters more than a few tokens per second, so buy for comfortable if you expect to.
06

See what else that card runs

Each card page repeats this sweep for every model we hold, answering what else the hardware is good for beyond Transformer (Adaptive Input Embeddings) WT103.

Answers

Transformer (Adaptive Input Embeddings) WT103 — common questions

Transformer (Adaptive Input Embeddings) WT103— can I run it on a GPU holding 12 GB?

Yes. The card GeForce RTX 3080 Ti, holding 12 GB, runs it at a compression of Q8_0, using about 1.0 GB and generating roughly 1,564 tokens per second. The fit is comfortable.

Transformer (Adaptive Input Embeddings) WT103— can I run it on a GPU holding 16 GB?

Yes. The card Tesla V100 SXM2 16 GB, holding 16 GB, runs it at a compression of Q8_0, using about 1.0 GB and generating roughly 1,938 tokens per second. The fit is comfortable.

Transformer (Adaptive Input Embeddings) WT103— can I run it on a GPU holding 24 GB?

Yes. The card GeForce RTX 5090 D V2, holding 24 GB, runs it at a compression of Q8_0, using about 1.0 GB and generating roughly 2,298 tokens per second. The fit is comfortable.

Transformer (Adaptive Input Embeddings) WT103— is it open source?

Its weights are published, so it 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.

Transformer (Adaptive Input Embeddings) WT103— how many parameters does it have?

It has a parameter count of 247M. Table 2. 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.

Transformer (Adaptive Input Embeddings) WT103— who created it?

It was published by Facebook AI Research, based in United States of America, an organisation categorised as industry.

Transformer (Adaptive Input Embeddings) WT103— when was it released?

It was published in September 2018. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

Transformer (Adaptive Input Embeddings) WT103— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language modeling. These are the areas it was designed around; they describe intent rather than a hard boundary.

Transformer (Adaptive Input Embeddings) WT103— where can I download it?

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

Transformer (Adaptive Input Embeddings) WT103— how much compute was used to train it?

Training consumed around 4.5 × 10¹⁹ FLOP, on hardware recorded as 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.

Transformer (Adaptive Input Embeddings) WT103— can I run it 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 model is rarely worth using. Every figure here assumes the whole model is resident on the card.

Transformer (Adaptive Input Embeddings) WT103— would two GPUs run it faster?

Two cards buy memory rather than speed, which matters only if one card cannot hold it. The number that can: 818. So a second card is rarely the answer here.

Transformer (Adaptive Input Embeddings) WT103— why does the quantisation differ between cards?

Because capacity varies, so does how hard it has to be squeezed. The number of distinct levels in the table above: 1. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

Transformer (Adaptive Input Embeddings) WT103— how accurate are these speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked, which is why each is published as a range rather than a single number. One example: 8,231–21,948 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

Transformer (Adaptive Input Embeddings) WT103— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Tesla C1080, with a memory capacity of 4 GB. It runs the model at a compression of Q8_0 using about 1.0 GB, and produces roughly 149 tokens per second. The number of cards able to run it in total: 818.

Transformer (Adaptive Input Embeddings) WT103— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 13,718 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and the number of cards clearing that: 818.

Transformer (Adaptive Input Embeddings) WT103— how much VRAM does it need?

It needs about 1.0 GB at a compression of Q8_0, 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.

Transformer (Adaptive Input Embeddings) WT103— can I run it on a GPU holding 8 GB?

Yes. The card CMP 170HX 8 GB, holding 8 GB, runs it at a compression of Q8_0, using about 1.0 GB and generating roughly 2,555 tokens per second. The fit is comfortable.

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