Papers › MCUFormer: Deploying Vision Transformers on Microcontrollers with Limited Memory

MCUFormer: Deploying Vision Transformers on Microcontrollers with Limited Memory

25 Oct 2023NeurIPS 2023 11arXiv:2310.16898archive 2025-07-28

Yinan Liang, Ziwei Wang, Xiuwei Xu, Yansong Tang, Jie zhou, Jiwen Lu

Due to the high price and heavy energy consumption of GPUs, deploying deep models on IoT devices such as microcontrollers makes significant contributions for ecological AI. Conventional methods successfully enable convolutional neural network inference of high resolution images on microcontrollers, while the framework for vision transformers that achieve the state-of-the-art performance in many vision applications still remains unexplored. In this paper, we propose a hardware-algorithm co-optimizations method called MCUFormer to deploy vision transformers on microcontrollers with extremely limited memory, where we jointly design transformer architecture and construct the inference operator library to fit the memory resource constraint. More specifically, we generalize the one-shot network architecture search (NAS) to discover the optimal architecture with highest task performance given the memory budget from the microcontrollers, where we enlarge the existing search space of vision transformers by considering the low-rank decomposition dimensions and patch resolution for memory reduction. For the construction of the inference operator library of vision transformers, we schedule the memory buffer during inference through operator integration, patch embedding decomposition, and token overwriting, allowing the memory buffer to be fully utilized to adapt to the forward pass of the vision transformer. Experimental results demonstrate that our MCUFormer achieves 73.62% top-1 accuracy on ImageNet for image classification with 320KB memory on STM32F746 microcontroller. Code is available at https://github.com/liangyn22/MCUFormer.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 14 code samples harvested from 1 repository linked to this paper: 10 ran (of which 4 constructed an object rather than computing a result; 6 with no instrument failure: 2 honoured, 0 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 4 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 14 samples from 1 repository, 10 ran (6 with no instrument failure; 4 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 10 samples that ran has no recorded date.

liangyn22/mcuformer officialmentioned in paperpytorchSyntology: 14 samples, 10 ran (4 instrument-failed) · 4 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

14 samples harvested: 10 ran (of which 4 constructed an object rather than computing a result; 6 with no instrument failure: 2 honoured, 0 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 4 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

2ran · honoured contract
4ran · our draft was wrong
4ran
4unverified

Licence: 14 of the 14 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from liangyn22/mcuformer. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

liangyn22/mcuformer official repository 14 samples, 10 ran (4 instrument-failed) · 4 unverified
LinearSuper model/supernet_transformer.py ran · metamorphic tier: deterministic no licence file found · pointer only · 46ae2d4f3fc0bdd6 · report fixture: returned LinearSuper · level recorded 2026-09-22
PatchembedSuper model/supernet_transformer.py ran · metamorphic tier: deterministic no licence file found · pointer only · a137a6efb264aa07 · report fixture: returned PatchembedSuper · level recorded 2026-09-22
RelativePosition2D_super model/supernet_transformer.py ran · metamorphic tier: deterministic no licence file found · pointer only · 701ac4868b3a9368 · report fixture: returned RelativePosition2D_super · level recorded 2026-09-22
_ntuple in 17 papers model/supernet_transformer.py ran · our draft was wrong no licence file found · pointer only · e2c1a7dffcdbe918 · report fixture: signature n:int/0/int · returned function · level recorded 2026-09-22
calc_dropout model/supernet_transformer.py ran · honoured contract fingerprinted no licence file found · pointer only · 72a860a9aff50105 · report fixture: signature dropout:float/0/float, sample_embed_dim:int/0/int, super_embed_dim:int/0/int · returned float · level recorded 2026-09-22
decode_cand_tuple evolution.py ran · our draft was wrong no licence file found · pointer only · b328ce4411979624 · report fixture: signature cand_tuple:tuple/1/int · returned tuple · level recorded, date not recorded
gelu in 308 papers
census: 43 implementations from 74 papers, 6 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 2, kind float, dtype float32
model/supernet_transformer.py ran · honoured contract fingerprinted no licence file found · pointer only · 08a5be1ffa0676e3 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
qkv_super model/supernet_transformer.py ran · metamorphic tier: deterministic no licence file found · pointer only · c356588a20cf831d · report fixture: returned qkv_super · level recorded 2026-09-22
sample_bias model/supernet_transformer.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 963bc33b77cf8148 · report fixture: signature bias:Tensor/1/float, sample_out_dim:int/0/int · returned Tensor · level recorded 2026-09-22
sample_weight in 8 papers model/supernet_transformer.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 1d3488a4b38ea75d · report fixture: signature weight:Tensor/2/float, sample_in_dim:int/0/int, sample_out_dim:int/0/int · returned Tensor · level recorded 2026-09-22
AttentionSuper model/supernet_transformer.py unverified no licence file found · pointer only · 765f54d24658a76a · report
LayerNormSuper model/supernet_transformer.py unverified no licence file found · pointer only · e4f3e6207e219ef8 · report
TransformerEncoderLayer in 90 papers model/supernet_transformer.py unverified no licence file found · pointer only · a115d422c7fbcbe2 · report
Vision_TransformerSuper model/supernet_transformer.py unverified no licence file found · pointer only · 09a88c3ccde67487 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for MCUFormer (graph method: MCUFormer). 17 bricks; individually on a fixture: 13 ran (of which 5 constructed an object rather than computing a result; 7 with no instrument failure: 2 honoured, 0 violated, 5 with no contract checked; 6 where Syntology's instrument failed) · 4 unverified. The assembly has not been run end-to-end.

  1. Vision_TransformerSuper entry · level 0 · unverified · 09a88c3ccde67487
  2. LayerNormSuper helper · level 0 · unverified · e4f3e6207e219ef8
  3. LinearSuper helper · level 1 · ran · 46ae2d4f3fc0bdd6
  4. PatchembedSuper model · level 1 · ran · a137a6efb264aa07
  5. TransformerEncoderLayer model · level 0 · unverified · a115d422c7fbcbe2
  6. calc_dropout helper · level 1 · ran · honoured contract · 72a860a9aff50105
  7. trunc_normal_ helper · level 1 · ran; the contract check could not run · 605878f6dcef72e9
  8. sample_bias helper · level 1 · ran · our draft was wrong · 963bc33b77cf8148
  9. sample_weight helper · level 1 · ran · our draft was wrong · 1d3488a4b38ea75d
  10. AttentionSuper model · level 0 · unverified · 765f54d24658a76a
  11. DropPath helper · level 1 · ran · 691dde70dfd8362b
  12. drop_path helper · level 1 · ran; the contract check could not run · 39eace7e2822504f
  13. gelu helper · level 1 · ran · honoured contract · 08a5be1ffa0676e3
  14. _no_grad_trunc_normal_ helper · level 1 · ran; the contract check could not run · 813ed3a96c17cf42
  15. RelativePosition2D_super helper · level 1 · ran · 701ac4868b3a9368
  16. qkv_super model · level 1 · ran · c356588a20cf831d
  17. _ntuple helper · level 1 · ran · our draft was wrong · e2c1a7dffcdbe918

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2310.16898") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Image Classificationimage-classification

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections