{"id":182824,"date":"2026-08-11T04:54:15","date_gmt":"2026-08-11T02:54:15","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3"},"modified":"2026-08-12T18:21:41","modified_gmt":"2026-08-12T16:21:41","slug":"yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","status":"publish","type":"post","link":"https:\/\/prohoster.info\/hr\/blog\/news\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","title":{"rendered":"Jezik model s 28,9 milijuna parametara radi na ESP32-S3","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

<\/i><\/a>
\nRazvija\u010d Via\u010deslav Serbov, poznat pod pseudonimom slvDev, prikazao je<\/a> potpuno lokalnu generaciju teksta na mikrocontrolleru ESP32-S3. Njegov je jezikovni model stvoren s 28,9 milijuna parametara i isporu\u010duje oko 9,9 tokena u sekundi, bez pozivanja na poslu\u017eitelj ili druge vanjske ra\u010dunalne resurse. Generirane kratke pri\u010de prikazuju se na priklju\u010denom OLED zaslonu.<\/p>\n

 <\/p>\n

Projekt esp32-ai<\/a> testiran je na modulu ESP32-S3 N16R8, koji ima 512 KB ugra\u0111ene SRAM, 8 MB PSRAM i 16 MB flash memorije. Kvantificirani model do \u010detiri bita zauzima 14,9 MB. U repozitoriju su objavljeni izvorni kod izvr\u0161ne okoline na C-u, sredstva za obuku i kvantizaciju modela, firmware, testovi i skripte za u\u010ditavanje na plo\u010du.<\/p>\n

 <\/p>\n

Navedene 28,9 milijuna parametara ne treba izravno uspore\u0111ivati s brojem parametara obi\u010dnih desktop LLM. Prema detaljnom izvje\u0161taju razvija\u010da,<\/a>model se sastoji od gusto procesorskog jezgra sa otprilike 559 tisu\u0107a parametara, izlaznog sloja sa 3,1 milijuna parametara i tablice slo\u017eenih vektorskih prikaza od otprilike 25 milijuna parametara. To\u010dno posljednje osigurava najve\u0107i dio formalne veli\u010dine modela, ali se ne obra\u0111uje u cijelosti prilikom generacije svakog tokena.<\/p>\n

 <\/p>\n

Za smje\u0161taj modela koriste se tri razine memorije. Naj\u010de\u0161\u0107e kori\u0161teni podaci nalaze se u unutarnjoj SRAM, izlazni sloj, KV-cache i privremeni baferi nalaze se u PSRAM-u, a tablica od 25 milijuna ostaje u flash memoriji i dostupna je putem mapiranja u adresnom prostoru. Za svaki token se \u010dita samo oko \u0161est redaka ukupne veli\u010dine otprilike 450 bajtova.<\/p>\n

 <\/p>\n

Ova naprava temelji se na tehnologiji Per-Layer Embeddings, koja je primijenjena od strane Googlea u Gemma 3n. Prema dokumentaciji Googlea,<\/a>PLE-parametri se mogu pohraniti izvan radne memorije modela i dodati u proces izlaza dok se obra\u0111uju pojedini slojevi. U esp32-ai ovaj pristup je prenesen na hijerarhiju memorije mikrocontrollera, gdje je brza SRAM posebno ograni\u010dena.<\/p>\n

 <\/p>\n

Prva ispravna verzija izvr\u0161nog okru\u017eenja napisanog u C-u isporu\u010divala je samo 0,57 tokena u sekundi. Nakon premje\u0161tanja izlaznog sloja u PSRAM, prelaska na osmobitne aktivacije, raspodjele izra\u010duna izme\u0111u dva jezgre Xtensa LX7 i drugih optimizacija, ka\u0161njenje je smanjeno na 94,9 ms po koraku modela. Mjerena brzina na kraju je dostigla 9,88 tokena u sekundi uzimaju\u0107i u obzir cijeli proces generacije. Glavno ograni\u010denje sada je propusnost PSRAM-a pri \u010ditanju izlaznog sloja.<\/p>\n

 <\/p>\n

Model je treniran na sinteti\u010dkom skupu podataka TinyStories<\/a>, koji se sastoji od jednostavnih engleskih pri\u010da s ograni\u010denim rje\u010dnikom. Stoga mo\u017ee nastaviti fraze i skladati kratke koherentne pri\u010de, ali nije namijenjena za odgovaranje na pitanja, izvr\u0161avanje instrukcija, programiranje ili reprodukciju stvarnog znanja. Razvija\u010d ovaj rad prvenstveno smatra demonstracijom u\u010dinkovite implementacije modela u memoriji mikrokontrolera, a ne kao spremnog autonomnog chat-bota.<\/p>\n

 <\/p>\n

U a\u017euriranoj dokumentaciji autor posebno nagla\u0161ava da je esp32-ai samostalni razvoj. Projekti llama2.c<\/a> Andreyja Karpatija i ranija verzija 260.000 modela na ESP32-S3 navedeni su samo kao smjernice i primjeri prethodnih radova: kod, kontrolne to\u010dke i metodologija nisu preuzeti iz njih. Izvorni kod esp32-ai distribuira se pod MIT licencom<\/a>.<\/p>\n

 <\/p>\n

Izvor: linux.org.ru<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

\u0420\u0430\u0437\u0440\u0430\u0431\u043e\u0442\u0447\u0438\u043a \u0412\u044f\u0447\u0435\u0441\u043b\u0430\u0432 \u0421\u0435\u0440\u0431\u043e\u0432, \u0432\u044b\u0441\u0442\u0443\u043f\u0430\u044e\u0449\u0438\u0439 \u043f\u043e\u0434 \u043f\u0441\u0435\u0432\u0434\u043e\u043d\u0438\u043c\u043e\u043c slvDev, \u043f\u0440\u043e\u0434\u0435\u043c\u043e\u043d\u0441\u0442\u0440\u0438\u0440\u043e\u0432\u0430\u043b \u043f\u043e\u043b\u043d\u043e\u0441\u0442\u044c\u044e \u043b\u043e\u043a\u0430\u043b\u044c\u043d\u0443\u044e \u0433\u0435\u043d\u0435\u0440\u0430\u0446\u0438\u044e \u0442\u0435\u043a\u0441\u0442\u0430 \u043d\u0430 \u043c\u0438\u043a\u0440\u043e\u043a\u043e\u043d\u0442\u0440\u043e\u043b\u043b\u0435\u0440\u0435 ESP32-S3. \u0421\u043e\u0437\u0434\u0430\u043d\u043d\u0430\u044f \u0438\u043c \u044f\u0437\u044b\u043a\u043e\u0432\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c \u0441\u043e\u0434\u0435\u0440\u0436\u0438\u0442 28,9 \u043c\u043b\u043d \u043f\u0430\u0440\u0430\u043c\u0435\u0442\u0440\u043e\u0432 \u0438 \u0432\u044b\u0434\u0430\u0451\u0442 \u043e\u043a\u043e\u043b\u043e 9,9 \u0442\u043e\u043a\u0435\u043d\u0430 \u0432 \u0441\u0435\u043a\u0443\u043d\u0434\u0443, \u043d\u0435 \u043e\u0431\u0440\u0430\u0449\u0430\u044f\u0441\u044c \u043a \u0441\u0435\u0440\u0432\u0435\u0440\u0443 \u0438\u043b\u0438 \u0434\u0440\u0443\u0433\u0438\u043c \u0432\u043d\u0435\u0448\u043d\u0438\u043c \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u043c \u0440\u0435\u0441\u0443\u0440\u0441\u0430\u043c. \u0421\u0433\u0435\u043d\u0435\u0440\u0438\u0440\u043e\u0432\u0430\u043d\u043d\u044b\u0435 \u043a\u043e\u0440\u043e\u0442\u043a\u0438\u0435 \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u044b \u0432\u044b\u0432\u043e\u0434\u044f\u0442\u0441\u044f \u043d\u0430 \u043f\u043e\u0434\u043a\u043b\u044e\u0447\u0451\u043d\u043d\u044b\u0439 OLED-\u0434\u0438\u0441\u043f\u043b\u0435\u0439.   \u041f\u0440\u043e\u0435\u043a\u0442 esp32-ai \u0438\u0441\u043f\u044b\u0442\u0430\u043d \u043d\u0430 \u043c\u043e\u0434\u0443\u043b\u0435 ESP32-S3 N16R8, \u0440\u0430\u0441\u043f\u043e\u043b\u0430\u0433\u0430\u044e\u0449\u0435\u043c […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":9,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-182824","post","type-post","status-publish","format-standard","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t\n\t\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t\n\n","aioseo_head_json":{"title":"\ud83e\udd47Jezik model s 28,9 milijuna parametara radi na ESP32-S3 | ProHoster","description":".","canonical_url":"https:\/\/prohoster.info\/hr\/blog\/news\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"hr_HR","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u042f\u0437\u044b\u043a\u043e\u0432\u0430\u044f \u043c\u043e\u0434\u0435\u043b\u044c \u0441 28,9 \u043c\u043b\u043d \u043f\u0430\u0440\u0430\u043c\u0435\u0442\u0440\u043e\u0432 \u0437\u0430\u0440\u0430\u0431\u043e\u0442\u0430\u043b\u0430 \u043d\u0430 ESP32-S3 | ProHoster","og:description":".","og:url":"https:\/\/prohoster.info\/hr\/blog\/news\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2026-08-11T02:54:15+00:00","article:modified_time":"2026-08-12T16:21:41+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"182824","title":null,"description":null,"keywords":null,"keyphrases":{"focus":[],"additional":[]},"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":"default","schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":null,"breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2026-08-11 06:43:49","updated":"2026-08-11 06:43:49","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts\/182824","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/comments?post=182824"}],"version-history":[{"count":1,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts\/182824\/revisions"}],"predecessor-version":[{"id":182897,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts\/182824\/revisions\/182897"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/media?parent=182824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/categories?post=182824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/tags?post=182824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}