{"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\/sv\/blog\/news\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","title":{"rendered":"Spr\u00e5kmodellen med 28,9 miljoner parametrar fungerar p\u00e5 ESP32-S3","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

<\/i><\/a>
\nUtvecklaren Vyacheslav Serbov, som g\u00e5r under pseudonymen slvDev, visade<\/a> en helt lokal textgenerering p\u00e5 mikrokontrollern ESP32-S3. Den spr\u00e5kmodell han skapade inneh\u00e5ller 28,9 miljoner parametrar och producerar cirka 9,9 token per sekund, utan att beh\u00f6va kontakta en server eller andra externa ber\u00e4kningsresurser. De genererade korta ber\u00e4ttelserna visas p\u00e5 den anslutna OLED-sk\u00e4rmen.<\/p>\n

 <\/p>\n

Projektet esp32-ai<\/a> testades p\u00e5 modulen ESP32-S3 N16R8, som har 512 KB inbyggd SRAM, 8 MB PSRAM och 16 MB flash-minne. Den kvantifierade modellen p\u00e5 fyra bitar tar upp 14,9 MB. I arkivet har k\u00e4llkoden f\u00f6r k\u00f6rmilj\u00f6n i C, verktyg f\u00f6r tr\u00e4ning och kvantisering av modellen, firmware, tester och skript f\u00f6r dess uppladdning till kortet publicerats.<\/p>\n

 <\/p>\n

De angivna 28,9 miljonerna parametrar b\u00f6r inte j\u00e4mf\u00f6ras direkt med antalet parametrar i vanliga station\u00e4ra LLM. Enligt den detaljerade rapporten fr\u00e5n utvecklaren<\/a>, best\u00e5r modellen av en t\u00e4t ber\u00e4kningsk\u00e4rna med cirka 559 tusen parametrar, ett utmatningsskikt med 3,1 miljoner parametrar och en tabell med lagrade vektorrepr\u00e4sentationer p\u00e5 cirka 25 miljoner parametrar. Just den sista ansvarar f\u00f6r den st\u00f6rsta delen av modellens formella storlek, men behandlas inte helt vid generering av varje token.<\/p>\n

 <\/p>\n

F\u00f6r att rymma modellen anv\u00e4nds tre niv\u00e5er av minne. De mest ofta anv\u00e4nda uppgifterna finns i den interna SRAM, utmatningsskiktet, KV-cache och tempor\u00e4ra buffrar finns i PSRAM, medan den 25-miljoners tabellen finns i flash-minnet och \u00e4r tillg\u00e4nglig genom avbildning i adressutrymmet. F\u00f6r varje token l\u00e4ses endast cirka sex rader, med en total volym p\u00e5 cirka 450 byte.<\/p>\n

 <\/p>\n

Denna enhet \u00e4r baserad p\u00e5 teknologin Per-Layer Embeddings, som Google har till\u00e4mpat i Gemma 3n. Enligt Google-dokumentationen<\/a>, kan PLE-parametrarna lagras utanf\u00f6r modellens arbetsminne och l\u00e4ggas till i utdata processen allteftersom individuella lager bearbetas. I esp32-ai har denna metod \u00f6verf\u00f6rts till mikrokontrollerns minneshierarki, d\u00e4r snabb SRAM \u00e4r s\u00e4rskilt begr\u00e4nsad.<\/p>\n

 <\/p>\n

Den f\u00f6rsta korrekta versionen av den C-baserade k\u00f6rmilj\u00f6n producerade endast 0,57 token per sekund. Efter att ha placerat utg\u00e5ngslagret i PSRAM, \u00f6verg\u00e5tt till \u00e5ttabitiga aktiveringar, distribuerat ber\u00e4kningar mellan tv\u00e5 Xtensa LX7-k\u00e4rnor och genomf\u00f6rt andra optimeringar reducerades f\u00f6rdr\u00f6jningen till 94,9 ms per modellsteg. Den slutgiltiga uppm\u00e4tta hastigheten n\u00e5dde 9,88 token per sekund med h\u00e4nsyn till hela genereringsprocessen. Den huvudsakliga begr\u00e4nsningen \u00e4r nu PSRAM:s bandbredd vid l\u00e4sning av utg\u00e5ngslagret.<\/p>\n

 <\/p>\n

Modellen \u00e4r tr\u00e4nad p\u00e5 ett syntetiskt dataset TinyStories<\/a>, best\u00e5ende av enkla engelskspr\u00e5kiga ber\u00e4ttelser med ett begr\u00e4nsat ordf\u00f6rr\u00e5d. D\u00e4rf\u00f6r kan den forts\u00e4tta meningar och skapa korta sammanh\u00e4ngande historier, men den \u00e4r inte avsedd f\u00f6r att svara p\u00e5 fr\u00e5gor, utf\u00f6ra instruktioner, programmera eller \u00e5terge faktiska kunskaper. Utvecklaren ser arbetet fr\u00e4mst som en demonstration av effektiv modellplacering i mikrokontrollerens minne, snarare \u00e4n som en f\u00e4rdig autonom chattbot.<\/p>\n

 <\/p>\n

I den aktuella dokumentationen p\u00e5pekar f\u00f6rfattaren s\u00e4rskilt att esp32-ai \u00e4r en sj\u00e4lvst\u00e4ndig utveckling. Projekten llama2.c<\/a> Andrej Karpatys och en tidigare version av en 260 000-parametrars modell p\u00e5 ESP32-S3 n\u00e4mns endast som referenser och exempel p\u00e5 tidigare arbeten: kod, punktkontroller och metodik har inte direkt l\u00e5nats fr\u00e5n dem. K\u00e4llkoden f\u00f6r esp32-ai sprids under MIT-licensen<\/a>.<\/p>\n

 <\/p>\n

K\u00e4lla: 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\udd47Spr\u00e5kmodellen med 28,9 miljoner parametrar k\u00f6rdes p\u00e5 ESP32-S3 | ProHoster","description":".","canonical_url":"https:\/\/prohoster.info\/sv\/blog\/news\/yazykovaya-model-s-289-mln-parametrov-zarabotala-na-esp32-s3","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"sv_SE","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\/sv\/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\/sv\/wp-json\/wp\/v2\/posts\/182824","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/users\/9"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/comments?post=182824"}],"version-history":[{"count":1,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts\/182824\/revisions"}],"predecessor-version":[{"id":182897,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/posts\/182824\/revisions\/182897"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/media?parent=182824"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/categories?post=182824"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/sv\/wp-json\/wp\/v2\/tags?post=182824"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}