{"id":105165,"date":"2022-09-25T21:36:38","date_gmt":"2022-09-25T19:36:39","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/otkryt-kod-sistemy-raspoznavaniya-i-perevoda-rechi-whisper"},"modified":"2022-09-25T21:36:38","modified_gmt":"2022-09-25T19:36:39","slug":"otkryt-kod-sistemy-raspoznavaniya-i-perevoda-rechi-whisper","status":"publish","type":"post","link":"https:\/\/prohoster.info\/cs\/blog\/news\/otkryt-kod-sistemy-raspoznavaniya-i-perevoda-rechi-whisper","title":{"rendered":"Byl otev\u0159en k\u00f3d syst\u00e9mu rozpozn\u00e1v\u00e1n\u00ed a p\u0159ekladu \u0159e\u010di Whisper","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

Projekt OpenAI, kter\u00fd se zab\u00fdv\u00e1 rozvojem ve\u0159ejn\u011b dostupn\u00fdch projekt\u016f v oblasti um\u011bl\u00e9 inteligence, zve\u0159ejnil poznatky spojen\u00e9 se syst\u00e9mem rozpozn\u00e1v\u00e1n\u00ed \u0159e\u010di Whisper. Tvrd\u00ed se, \u017ee pro anglickou \u0159e\u010d syst\u00e9m poskytuje \u00farovn\u011b spolehlivosti a p\u0159esnosti automatick\u00e9ho rozpozn\u00e1v\u00e1n\u00ed bl\u00edzk\u00e9 rozpozn\u00e1n\u00ed \u010dlov\u011bkem. K\u00f3d referen\u010dn\u00ed implementace zalo\u017een\u00e9 na frameworku PyTorch a sada ji\u017e vycvi\u010den\u00fdch model\u016f jsou nyn\u00ed k dispozici. K\u00f3d je otev\u0159en pod licenc\u00ed MIT. <\/p>\n

Pro tr\u00e9nink modelu bylo pou\u017eito 680 000 hodin \u0159e\u010dov\u00fdch dat, shrom\u00e1\u017ed\u011bn\u00fdch z n\u011bkolika sb\u00edrek pokr\u00fdvaj\u00edc\u00edch r\u016fzn\u00e9 jazyky a tematick\u00e9 oblasti. P\u0159ibli\u017en\u011b 1\/3 pou\u017eit\u00fdch \u0159e\u010dov\u00fdch dat p\u0159i tr\u00e9nov\u00e1n\u00ed poch\u00e1z\u00ed z jazyk\u016f odli\u0161n\u00fdch od angli\u010dtiny. Navrhovan\u00fd syst\u00e9m spr\u00e1vn\u011b zpracov\u00e1v\u00e1 situace, jako je v\u00fdslovnost s p\u0159\u00edzvukem, p\u0159\u00edtomnost pozad\u00ed hluku a pou\u017eit\u00ed technick\u00e9ho \u017eargonu. Krom\u011b p\u0159episu \u0159e\u010di do textu m\u016f\u017ee syst\u00e9m tak\u00e9 p\u0159ekl\u00e1dat \u0159e\u010d z libovoln\u00e9ho jazyka do angli\u010dtiny a identifikovat p\u0159\u00edtomnost \u0159e\u010di v zvukov\u00e9m toku. <\/p>\n

Modely jsou vytvo\u0159eny ve dvou podob\u00e1ch: model pro anglick\u00fd jazyk a v\u00edcejazy\u010dn\u00fd model, kter\u00fd podporuje tak\u00e9 ru\u0161tinu, ukrajin\u0161tinu a b\u011bloru\u0161tinu. Ka\u017ed\u00e1 podoba se nav\u00edc d\u011bl\u00ed na 5 variant, kter\u00e9 se li\u0161\u00ed velikost\u00ed a po\u010dtem parametr\u016f zahrnut\u00fdch v modelu. \u010c\u00edm v\u011bt\u0161\u00ed velikost, t\u00edm v\u011bt\u0161\u00ed p\u0159esnost a kvalita rozpozn\u00e1n\u00ed, ale tak\u00e9 vy\u0161\u0161\u00ed po\u017eadavky na velikost videopam\u011bti GPU a ni\u017e\u0161\u00ed v\u00fdkon. Nap\u0159\u00edklad minim\u00e1ln\u00ed varianta zahrnuje 39 milion\u016f parametr\u016f a vy\u017eaduje 1 GB videopam\u011bti, zat\u00edmco maxim\u00e1ln\u00ed zahrnuje 1550 milion\u016f parametr\u016f a vy\u017eaduje 10 GB videopam\u011bti. Minim\u00e1ln\u00ed varianta je 32kr\u00e1t rychlej\u0161\u00ed ne\u017e maxim\u00e1ln\u00ed.

\"Byl<\/center> <\/p>\n

Syst\u00e9m vyu\u017e\u00edv\u00e1 architekturu neuronov\u00e9 s\u00edt\u011b \u201eTransformer\u201c, kter\u00e1 zahrnuje vz\u00e1jemn\u011b propojen\u00e9 enkod\u00e9r a dekod\u00e9r. Zvuk je rozd\u011blen na 30sekundov\u00e9 \u00faryvky, kter\u00e9 jsou p\u0159evedeny na log-Mel spektrogram a odesl\u00e1ny do enkod\u00e9ru. V\u00fdsledek pr\u00e1ce enkod\u00e9ru je p\u0159ed\u00e1n dekod\u00e9ru, kter\u00fd predikuje textovou reprezentaci sm\u00edchanou se speci\u00e1ln\u00edmi tokeny, kter\u00e9 umo\u017e\u0148uj\u00ed ve spole\u010dn\u00e9m modelu \u0159e\u0161it \u00fakoly jako je ur\u010den\u00ed jazyka, sledov\u00e1n\u00ed chronologie v\u00fdslovnosti fr\u00e1z\u00ed, p\u0159epis \u0159e\u010di v r\u016fzn\u00fdch jazyc\u00edch a p\u0159eklad do angli\u010dtiny.
\n
Zdroj: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"

\u041f\u0440\u043e\u0435\u043a\u0442 OpenAI, \u0437\u0430\u043d\u0438\u043c\u0430\u044e\u0449\u0438\u0439\u0441\u044f \u0440\u0430\u0437\u0432\u0438\u0442\u0438\u0435\u043c \u043e\u0431\u0449\u0435\u0434\u043e\u0441\u0442\u0443\u043f\u043d\u044b\u0445 \u043f\u0440\u043e\u0435\u043a\u0442\u043e\u0432 \u0432 \u043e\u0431\u043b\u0430\u0441\u0442\u0438 \u0438\u0441\u043a\u0443\u0441\u0441\u0442\u0432\u0435\u043d\u043d\u043e\u0433\u043e \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0430, \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043b \u043d\u0430\u0440\u0430\u0431\u043e\u0442\u043a\u0438, \u0441\u0432\u044f\u0437\u0430\u043d\u043d\u044b\u0435 \u0441 \u0441\u0438\u0441\u0442\u0435\u043c\u043e\u0439 \u0440\u0430\u0441\u043f\u043e\u0437\u043d\u0430\u0432\u0430\u043d\u0438\u044f \u0440\u0435\u0447\u0438 Whisper. \u0423\u0442\u0432\u0435\u0440\u0436\u0434\u0430\u0435\u0442\u0441\u044f, \u0447\u0442\u043e \u0434\u043b\u044f \u0440\u0435\u0447\u0438 \u043d\u0430 \u0430\u043d\u0433\u043b\u0438\u0439\u0441\u043a\u043e\u043c \u044f\u0437\u044b\u043a\u0435 \u0441\u0438\u0441\u0442\u0435\u043c\u0430 \u043e\u0431\u0435\u0441\u043f\u0435\u0447\u0438\u0432\u0430\u0435\u0442 \u0443\u0440\u043e\u0432\u043d\u0438 \u043d\u0430\u0434\u0451\u0436\u043d\u043e\u0441\u0442\u0438 \u0438 \u0442\u043e\u0447\u043d\u043e\u0441\u0442\u0438 \u0430\u0432\u0442\u043e\u043c\u0430\u0442\u0438\u0447\u0435\u0441\u043a\u043e\u0433\u043e \u0440\u0430\u0441\u043f\u043e\u0437\u043d\u0430\u0432\u0430\u043d\u0438\u044f \u0431\u043b\u0438\u0437\u043a\u0438\u0435 \u043a \u0440\u0430\u0441\u043f\u043e\u0437\u043d\u0430\u0432\u0430\u043d\u0438\u044e \u0447\u0435\u043b\u043e\u0432\u0435\u043a\u043e\u043c. \u041e\u0442\u043a\u0440\u044b\u0442\u044b \u043a\u043e\u0434 \u044d\u0442\u0430\u043b\u043e\u043d\u043d\u043e\u0439 \u0440\u0435\u0430\u043b\u0438\u0437\u0430\u0446\u0438\u0438 \u043d\u0430 \u0431\u0430\u0437\u0435 \u0444\u0440\u0435\u0439\u043c\u0432\u043e\u0440\u043a\u0430 PyTorch \u0438 \u043d\u0430\u0431\u043e\u0440 \u0443\u0436\u0435 \u043e\u0431\u0443\u0447\u0435\u043d\u043d\u044b\u0445 \u043c\u043e\u0434\u0435\u043b\u0435\u0439, \u0433\u043e\u0442\u043e\u0432\u044b\u0445 \u0434\u043b\u044f \u0438\u0441\u043f\u043e\u043b\u044c\u0437\u043e\u0432\u0430\u043d\u0438\u044f. \u041a\u043e\u0434 \u043e\u0442\u043a\u0440\u044b\u0442 […]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":105166,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-105165","post","type-post","status-publish","format-standard","has-post-thumbnail","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\t\t\n\n","aioseo_head_json":{"title":"\ud83e\udd47Otev\u0159en\u00fd k\u00f3d syst\u00e9mu rozpozn\u00e1v\u00e1n\u00ed a p\u0159eklad\u016f \u0159e\u010di Whisper | ProHoster","description":"Projekt OpenAI, kter\u00fd se zab\u00fdv\u00e1 v\u00fdvojem ve\u0159ejn\u011b dostupn\u00fdch projekt\u016f v oblasti um\u011bl\u00e9 inteligence, zve\u0159ejnil v\u00fdsledky souvisej\u00edc\u00ed se syst\u00e9mem rozpozn\u00e1v\u00e1n\u00ed \u0159e\u010di Whisper.","canonical_url":"https:\/\/prohoster.info\/cs\/blog\/news\/otkryt-kod-sistemy-raspoznavaniya-i-perevoda-rechi-whisper","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"cs_CZ","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\u041e\u0442\u043a\u0440\u044b\u0442 \u043a\u043e\u0434 \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u0440\u0430\u0441\u043f\u043e\u0437\u043d\u0430\u0432\u0430\u043d\u0438\u044f \u0438 \u043f\u0435\u0440\u0435\u0432\u043e\u0434\u0430 \u0440\u0435\u0447\u0438 Whisper | ProHoster","og:description":"\u041f\u0440\u043e\u0435\u043a\u0442 OpenAI, \u0437\u0430\u043d\u0438\u043c\u0430\u044e\u0449\u0438\u0439\u0441\u044f \u0440\u0430\u0437\u0432\u0438\u0442\u0438\u0435\u043c \u043e\u0431\u0449\u0435\u0434\u043e\u0441\u0442\u0443\u043f\u043d\u044b\u0445 \u043f\u0440\u043e\u0435\u043a\u0442\u043e\u0432 \u0432 \u043e\u0431\u043b\u0430\u0441\u0442\u0438 \u0438\u0441\u043a\u0443\u0441\u0441\u0442\u0432\u0435\u043d\u043d\u043e\u0433\u043e \u0438\u043d\u0442\u0435\u043b\u043b\u0435\u043a\u0442\u0430, \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043b \u043d\u0430\u0440\u0430\u0431\u043e\u0442\u043a\u0438, \u0441\u0432\u044f\u0437\u0430\u043d\u043d\u044b\u0435 \u0441 \u0441\u0438\u0441\u0442\u0435\u043c\u043e\u0439 \u0440\u0430\u0441\u043f\u043e\u0437\u043d\u0430\u0432\u0430\u043d\u0438\u044f \u0440\u0435\u0447\u0438 Whisper.","og:url":"https:\/\/prohoster.info\/cs\/blog\/news\/otkryt-kod-sistemy-raspoznavaniya-i-perevoda-rechi-whisper","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":"2022-09-25T19:36:39+00:00","article:modified_time":"2022-09-25T19:36:39+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":[],"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/105165","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/comments?post=105165"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/posts\/105165\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media\/105166"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/media?parent=105165"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/categories?post=105165"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/cs\/wp-json\/wp\/v2\/tags?post=105165"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}