Meta tags:

Headings (most frequently used words):

mobilenet, contents, features, see, also, references, external, links, v1, v2, v3, v4, v5,

Text of the page (most frequently used words):
the (40), and (25), 2024 (15), google (14), 2019 (14), 2017 (13), #mobilenet (12), edit (12), for (12), mobile (10), was (10), vision (10), convolution (9), this (8), computer (8), retrieved (8), 2025 (8), 2018 (8), models (8), arxiv (8), inverted (8), depthwise (8), wikipedia (7), neural (7), that (7), search (6), learning (6), 2023 (6), 2022 (6), mobilenetv3 (6), howard (6), andrew (6), chen (6), displaystyle (6), page (5), with (5), tensorflow (5), other (5), mobilenetv2 (5), june (5), published (5), they (5), add (4), contents (4), april (4), from (4), information (4), also (4), 2016 (4), networks (4), alphago (4), links (4), zhu (4), sandler (4), mark (4), network (4), cvpr (4), research (4), device (4), linear (4), bottlenecks (4), efficient (4), convolutional (4), times (4), which (4), number (4), channels (4), main (4), devices (4), designed (4), hide (4), move (4), sidebar (4), toggle (3), view (3), developers (3), terms (3), you (3), license (3), keras (3), pattern (3), recognition (3), computing (3), transformer (3), need (3), see (3), gemini (3), gemma (3), wang (3), weijun (3), adam (3), hartwig (3), architecture (3), ieee (3), deep (3), separable (3), convolutions (3), release (3), has (3), included (3), residual (3), two (3), small (3), latency (3), account (3), accuracy (3), activation (3), lower (3), input (3), each (3), features (3), family (3), tools (3), languages (2), table (2), contact (2), about (2), privacy (2), policy (2), available (2), commons (2), 2026 (2), categories (2), short (2), description (2), wikidata (2), software (2), machine (2), programming (2), learn (2), computational (2), differentiable (2), general (2), generative (2), attention (2), all (2), processing (2), jie (2), 2015 (2), fan (2), master (2), versions (2), brain (2), github (2), external (2), com (2), introducing (2), developer (2), one (2), marco (2), yang (2), mobilenetv4 (2), universal (2), eccv (2), squeeze (2), excitation (2), platform (2), aware (2), applications (2), tan (2), mingxing (2), pang (2), ruoming (2), vasudevan (2), vijay (2), quoc (2), 978 (2), isbn (2), 1109 (2), doi (2), 1807 (2), conference (2), liang (2), chieh (2), next (2), generation (2), menglong (2), residuals (2), mobilenets (2), references (2), lite (2), mobilenetv5 (2), series (2), bottleneck (2), special (2), publication (2), large (2), architectures (2), found (2), nas (2), were (2), into (2), swish (2), sigmoid (2), projection (2), layers (2), dimensional (2), structure (2), them (2), layer (2), reduces (2), during (2), mobilenetv1 (2), controls (2), smaller (2), faster (2), but (2), cost (2), resolution (2), multiplier (2), alpha (2), architectural (2), size (2), low (2), inference (2), appearance (2), upload (2), file (2), changes (2), history (2), read (2), article (2), log (2), create (2), donate (2), menu (2), topic, cookie, statement, statistics, code, conduct, legal, safety, contacts, disclaimers, text, under, additional, may, apply, using, site, agree, registered, trademark, non, profit, organization, wikimedia, foundation, inc, use, creative, attribution, sharealike, rendered, parsoid, last, edited, utc, hidden, different, articles, https, org, index, php, title, oldid, 1350591938, portals, technology, mindspore, flux, jax, theano, scikit, pytorch, libraries, spinnaker, memristor, vpu, tpu, ipu, hardware, inductive, bias, theory, ricci, calculus, neuromorphic, automatic, differentiation, statistical, manifold, geometry, category, workspace, labs, pre, trained, future, summit, antigravity, robotics, vids, notebooklm, dreambooth, veo, genie, videopoet, imagen, palm, chinchilla, 2021, lamda, xlnet, bert, nano, banana, sparrow, assistant, chatbots, tensor, unit, quantum, artificial, intelligence, lab, gato, efficientnet, wavenet, 2014, inception, alphagenome, alphaevolve, alphaproof, alphageometry, funsearch, alphadev, alphatensor, alphastar, alphafold, popular, culture, lee, sedol, hui, competitions, muzero, alphazero, zero, programs, deepmind, october, documentation, googleblog, guide, blog, shazeer, noam, fast, decoding, write, head, 1911, 02150, wightman, ross, huggingface, now, timm, qin, danfeng, leichner, chas, delakis, manolis, fornoni, luo, shixin, banbury, colby, chengxi, akin, berkin, aggarwal, vaibhav, tenghui, moro, daniele, ecosystem, 2404, 10518, shen, sun, gang, 7141, 7132, tien, zhang, xiao, alec, sze, vivienne, 300, 1804, 03230, 285, netadapt, adaptation, mnasnet, 2823, 7281, 3293, 00293, 11626, 2815, cvf, chu, grace, yukun, 1324, 1905, 02244, 1314, iccv, searching, november, mobi, zhmoginov, andrey, 1801, 04381, chollet, françois, 5386, 0457, 195, 1610, 02357, 1800, xception, open, source, kalenichenko, dmitry, weyand, tobias, andreetto, 1704, 04861, shortly, after, while, announcement, stated, technical, report, would, soon, not, yet, materialised, larger, than, largest, variant, inspired, unified, both, includes, these, cases, multi, query, transformers, september, mobilenetedgetpu, optimized, form, takes, achieve, good, trade, off, between, used, piecewise, approximations, called, modules, functions, pixel, removes, typical, relu, function, rationalized, arguing, nonlinear, loses, spaces, problematic, when, already, modify, traditional, block, instead, compressing, before, expansion, followed, then, back, down, helps, maintain, representational, capacity, allowing, operate, higher, feature, space, thus, preserving, more, flow, process, expand, march, uses, hyperparameters, values, lead, reduced, images, resolutions, result, potentially, rho, width, decomposes, single, standard, filters, channel, independently, pointwise, combines, outputs, factorization, significantly, its, innovation, incorporation, first, developed, laurent, sifre, internship, 2013, variation, improve, convergence, speed, model, alexnet, led, researchers, develop, five, improving, upon, previous, performance, efficiency, cnn, tasks, are, power, consumption, making, suitable, resource, constrained, like, originally, run, efficiently, embedded, systems, phones, edge, object, detection, image, classification, tree, slim, nets, repository, apache, python, written, stable, free, encyclopedia, item, projects, printable, version, download, pdf, print, export, switch, legacy, parser, get, shortened, url, cite, permanent, link, related, what, here, actions, english, talk, subsection, top, personal, pages, recent, community, portal, help, contribute, random, current, events, navigation, jump, content,

Text of the page (random words):
mobilenet wikipedia jump to content main menu main menu move to sidebar hide navigation main page contents current events random article about wikipedia contact us contribute help learn to edit community portal recent changes upload file special pages search search appearance donate create account log in personal tools donate create account log in contents move to sidebar hide top 1 features toggle features subsection 1 1 v1 1 2 v2 1 3 v3 1 4 v4 1 5 v5 2 see also 3 references 4 external links toggle the table of contents mobilenet add languages add links article talk english read edit view history tools tools move to sidebar hide actions read edit view history general what links here related changes upload file permanent link page information cite this page get shortened url switch to legacy parser print export download as pdf printable version in other projects wikidata item appearance move to sidebar hide from wikipedia the free encyclopedia family of computer vision models designed for efficient inference on mobile devices mobilenet developer google release april 2017 stable release v5 june 2025 written in python license apache license 2 0 repository github com tensorflow models tree master research slim nets mobilenet mobilenet is a family of convolutional neural network cnn architectures designed for image classification object detection and other computer vision tasks they are designed for small size low latency and low power consumption making them suitable for on device inference and edge computing on resource constrained devices like mobile phones and embedded systems they were originally designed to be run efficiently on mobile devices with tensorflow lite the need for efficient deep learning models on mobile devices led researchers at google to develop mobilenet as of june 2025 the family has five versions each improving upon the previous one in terms of performance and efficiency features edit v1 edit mobilenetv1 was published in april 2017 1 2 its main architectural innovation was incorporation of depthwise separable convolutions it was first developed by laurent sifre during an internship at google brain in 2013 as an architectural variation on alexnet to improve convergence speed and model size 3 the depthwise separable convolution decomposes a single standard convolution into two convolutions a depthwise convolution that filters each input channel independently and a pointwise convolution 1 1 displaystyle 1 times 1 convolution that combines the outputs of the depthwise convolution this factorization significantly reduces computational cost the mobilenetv1 has two hyperparameters a width multiplier α displaystyle alpha that controls the number of channels in each layer smaller values of α displaystyle alpha lead to smaller and faster models but at the cost of reduced accuracy and a resolution multiplier ρ displaystyle rho which controls the input resolution of the images lower resolutions result in faster processing but potentially lower accuracy v2 edit mobilenetv2 was published in march 2019 4 5 it uses inverted residual layers and linear bottlenecks inverted residuals modify the traditional residual block structure instead of compressing the input channels before the depthwise convolution they expand them this expansion is followed by a 1 1 displaystyle 1 times 1 depthwise convolution and then a 1 1 displaystyle 1 times 1 projection layer that reduces the number of channels back down this inverted structure helps to maintain representational capacity by allowing the depthwise convolution to operate on a higher dimensional feature space thus preserving more information flow during the convolutional process linear bottlenecks removes the typical relu activation function in the projection layers this was rationalized by arguing that that nonlinear activation loses information in lower dimensional spaces which is problematic when the number of channels is already small v3 edit mobilenetv3 was published in 2019 6 7 the publication included mobilenetv3 small mobilenetv3 large and mobilenetedgetpu optimized for pixel 4 they were found by a form of neural architecture search nas that takes mobile latency into account to achieve good trade off between accuracy and latency 8 9 it used piecewise linear approximations of swish and sigmoid activation functions which they called h swish and h sigmoid squeeze and excitation modules 10 and the inverted bottlenecks of mobilenetv2 v4 edit mobilenetv4 was published in september 2024 11 12 the publication included a large number of architectures found by nas inspired by vision transformers the v4 series included multi query attention 13 it also unified both inverted residual and inverted bottleneck from the v3 series with the universal inverted bottleneck which includes these two as special cases v5 edit mobilenetv5 s architecture was published shortly after the release of gemma 3n in june 2025 14 while the announcement stated a technical report on mobilenetv5 would be available soon this has not yet materialised the network is 10 times larger than the largest v4 variant 14 see also edit convolutional neural network deep learning tensorflow lite references edit howard andrew g zhu menglong chen bo kalenichenko dmitry wang weijun weyand tobias andreetto marco adam hartwig 2017 mobilenets efficient convolutional neural networks for mobile vision applications arxiv 1704 04861 cs cv mobilenets open source models for efficient on device vision research google june 14 2017 retrieved 2024 10 18 chollet françois 2017 xception deep learning with depthwise separable convolutions 2017 ieee conference on computer vision and pattern recognition cvpr pp 1800 1807 arxiv 1610 02357 doi 10 1109 cvpr 2017 195 isbn 978 1 5386 0457 1 sandler mark howard andrew zhu menglong zhmoginov andrey chen liang chieh 2018 mobilenetv2 inverted residuals and linear bottlenecks arxiv 1801 04381 cs cv mobilenetv2 the next generation of on device computer vision networks research google april 3 2018 retrieved 2024 10 18 introducing the next generation of on device vision models mobilenetv3 and mobi research google november 13 2019 retrieved 2024 10 18 howard andrew sandler mark chu grace chen liang chieh chen bo tan mingxing wang weijun zhu yukun pang ruoming vasudevan vijay le quoc v adam hartwig 2019 searching for mobilenetv3 iccv 2019 1314 1324 arxiv 1905 02244 tan mingxing chen bo pang ruoming vasudevan vijay sandler mark howard andrew le quoc v june 2019 mnasnet platform aware neural architecture search for mobile 2019 ieee cvf conference on computer vision and pattern recognition cvpr ieee pp 2815 2823 arxiv 1807 11626 doi 10 1109 cvpr 2019 00293 isbn 978 1 7281 3293 8 yang tien ju howard andrew chen bo zhang xiao go alec sandler mark sze vivienne adam hartwig 2018 netadapt platform aware neural network adaptation for mobile applications eccv 2018 285 300 arxiv 1804 03230 hu jie shen li sun gang 2018 squeeze and excitation networks eccv 2018 7132 7141 qin danfeng leichner chas delakis manolis fornoni marco luo shixin yang fan wang weijun banbury colby ye chengxi akin berkin aggarwal vaibhav zhu tenghui moro daniele howard andrew 2024 mobilenetv4 universal models for the mobile ecosystem arxiv 2404 10518 cs cv wightman ross mobilenet v4 now in timm huggingface co retrieved 2024 10 18 shazeer noam 2019 fast transformer decoding one write head is all you need arxiv 1911 02150 cs ne 1 2 introducing gemma 3n the developer guide google developers blog developers googleblog com retrieved 2026 04 10 external links edit mobilenet github retrieved 2024 10 18 keras documentation mobilenet mobilenetv2 and mobilenetv3 keras retrieved october 18 2024 v t e google ai google google brain google deepmind computer programs alphago versions alphago 2015 master 2016 alphago zero 2017 alphazero 2017 muzero 2019 competitions fan hui 2015 lee sedol 2016 ke jie 2017 in popular culture alphago 2017 other alphafold 2018 alphastar 2019 alphatensor 2022 alphadev 2023 funsearch 2023 alphageometry 2024 alphaproof 2024 alphaevolve 2025 alphagenome 2025 machine learning neural networks inception 2014 wavenet 2016 mobilenet 2017 transformer 2017 efficientnet 2019 gato 2022 other quantum artificial intelligence lab tensorflow tensor processing unit generative ai chatbots assistant 2016 sparrow 2022 gemini 2023 nano banana 2025 models bert 2018 xlnet 2019 t5 2019 lamda 2021 chinchilla 2022 palm 2022 imagen 2023 gemini 2023 videopoet 2024 gemma 2024 genie 2024 veo 2024 other dreambooth 2022 notebooklm 2023 vids 2024 gemini robotics 2025 antigravity 2025 see also attention is all you need future of go summit generative pre trained transformer google labs google workspace category commons v t e differentiable computing general differentiable programming information geometry statistical manifold automatic differentiation neuromorphic computing pattern recognition ricci calculus computational learning theory inductive bias hardware ipu tpu vpu memristor spinnaker software libraries tensorflow pytorch keras scikit learn theano jax flux jl mindspore portals computer programming technology retrieved from https en wikipedia org w index php title mobilenet oldid 1350591938 categories computer vision machine learning google software hidden categories articles with short description short description is different from wikidata this page was last edited on 22 april 2026 at 20 38 utc page was rendered with parsoid text is available under the creative commons attribution sharealike 4 0 license additional terms may apply by using this site you agree to the terms of use and privacy policy wikipedia is a registered trademark of the wikimedia foundation inc a non profit organization privacy policy about wikipedia disclaimers contact wikipedia legal safety contacts code of conduct developers statistics cookie statement mobile view search search toggle the table of contents mobilenet add languages add topic