TrOCR TPS calculator

Open weights Beihang University,Microsoft Research Asia 558M parameters September 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 · 66.1 tok/s

Fastest card

B200

6,072 tok/s · 180 GB

Which GPUs can run TrOCR?

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
6,072 tok/s 3,643–9,715 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.3 GB	Q8_0	Comfortable
6,072 tok/s 3,643–9,715 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.3 GB	Q8_0	Comfortable
4,849 tok/s 2,909–7,758 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.3 GB	Q8_0	Comfortable
4,849 tok/s 2,909–7,758 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.3 GB	Q8_0	Comfortable
3,878 tok/s 2,327–6,204 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
3,712 tok/s 2,227–5,939 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.3 GB	Q8_0	Comfortable
3,712 tok/s 2,227–5,939 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.3 GB	Q8_0	Comfortable
3,552 tok/s 2,131–5,683 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.3 GB	Q8_0	Comfortable
3,153 tok/s 1,892–5,044 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
3,153 tok/s 1,892–5,044 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
3,153 tok/s 1,892–5,044 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.3 GB	Q8_0	Comfortable
2,991 tok/s 1,794–4,785 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,550 tok/s 1,530–4,080 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,550 tok/s 1,530–4,080 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.3 GB	Q8_0	Comfortable
2,550 tok/s 1,530–4,080 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,550 tok/s 1,530–4,080 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
2,550 tok/s 1,530–4,080 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.3 GB	Q8_0	Comfortable
1,942 tok/s 1,165–3,107 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.3 GB	Q8_0	Comfortable
1,942 tok/s 1,165–3,107 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.3 GB	Q8_0	Comfortable
1,618 tok/s 971–2,589 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
1,584 tok/s 950–2,534 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.3 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.3 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.3 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.3 GB	Q8_0	Comfortable
1,548 tok/s 929–2,477 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.3 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: Beihang University,Microsoft Research Asia
Organisation type: Academia,Industry
Country: China
Published: 21 September 2021
Authors: Minghao Li, Tengchao Lv, Jingye Chen, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei

What it does

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

Domain: Vision
Task: Character recognition (OCR)

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: 558M
Training data: tokens

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: 32
Power draw: 19.4 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.

Why it is tracked: SOTA improvement
Record confidence: Confident
Citations: 616

Sources

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

Reference: TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run TrOCR

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 TrOCR

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

Hardware requirements in practice

Minimum card

Tesla C1080

Memory needed

1.3 GB

Fastest

6,072 tok/s

TrOCR is small enough at 558M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The entry point is the Tesla C1080: 4 GB of memory, Q8_0 compression, roughly 66.1 tokens per second.

At the other end, a B200 generates roughly 6,072 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

Where it came from

TrOCR was published by Beihang University,Microsoft Research Asia, in China, in September 2021. It comes out of academia,Industry.

It works in Vision, and is recorded as doing character recognition (OCR).

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

Understanding the speeds

Half the cards that hold it manage more than 170.5 tokens per second, and 809 exceed reading speed outright.

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.

Memory here is estimated from size rather than computed from the architecture, which is not recorded for this model — the numbers are indicative rather than exact.

How it was trained

The reason it appears in this catalogue at all is sOTA improvement.

Step by step

How to choose a GPU for TrOCR

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

Every card here has been checked against TrOCR — around 1.3 GB at Q8_0. Capacity is the gate — a card either holds it or it does not.
02

Decide how long your conversations run

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for TrOCR.
03

Choose how far you will compress it

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage TrOCR by squeezing it further than you would want.
04

Rank by throughput rather than spec sheet

Ranking by tokens per second for TrOCR follows memory bandwidth, not core counts, which is why the B200 tops it at 6,072 tok/s.
05

Check the fit verdict before buying

The fit column separates cards that just manage TrOCR from those with room to spare. Buy for the second if the context might grow.
06

Check the card from the other side

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond TrOCR.

Answers

TrOCR — common questions

How many parameters does TrOCR have?

TrOCR has 558M parameters. 558M table 5. 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 TrOCR?

TrOCR was published by Beihang University,Microsoft Research Asia, based in China, categorised as academia,Industry.

When was TrOCR released?

TrOCR was published in September 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 TrOCR used for?

TrOCR works in Vision, and is recorded as handling character recognition (OCR). Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download TrOCR?

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

Can I run TrOCR if it does not fit in my GPU?

It can be split between the card and system memory, but TrOCR generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run TrOCR faster?

Two cards buy memory rather than speed. That matters for TrOCR only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for TrOCR?

Because capacity varies, so does how hard TrOCR has to be squeezed — 1 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these TrOCR speed estimates?

These are estimates with real error bars. The fastest result here, 3,643–9,715 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

What GPU do I need to run TrOCR?

The smallest card in our catalogue that holds TrOCR is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.3 GB, and produces roughly 66.1 tokens per second. 818 cards in total can run it.

How fast is TrOCR on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 6,072 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run TrOCR clear that.

How much VRAM does TrOCR need?

About 1.3 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 TrOCR on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.3 GB and generating roughly 1,131 tokens per second — a comfortable fit.

Can I run TrOCR on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.3 GB and generating roughly 693 tokens per second — a comfortable fit.

Can I run TrOCR on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.3 GB and generating roughly 858 tokens per second — a comfortable fit.

Can I run TrOCR on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.3 GB and generating roughly 1,017 tokens per second — a comfortable fit.

Is TrOCR open source?

Its weights are published, so TrOCR 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.

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