{"id":157425,"date":"2026-02-07T17:12:29","date_gmt":"2026-02-07T15:12:29","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/inicziativa-po-vstraivaniyu-modelej-mashinnogo-obucheniya-v-yadro-linux"},"modified":"2026-02-07T17:12:29","modified_gmt":"2026-02-07T15:12:29","slug":"inicziativa-po-vstraivaniyu-modelej-mashinnogo-obucheniya-v-yadro-linux","status":"publish","type":"post","link":"https:\/\/prohoster.info\/sq\/blog\/novosti-interneta\/inicziativa-po-vstraivaniyu-modelej-mashinnogo-obucheniya-v-yadro-linux","title":{"rendered":"Iniciativa p\u00ebr integrimin e modeleve t\u00eb m\u00ebsimit t\u00eb makineris\u00eb n\u00eb thelbin e Linux","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
Vyacheslav Dubeyko from IBM initiated a discussion in the Linux kernel developers mailing list about using machine learning models within the kernel. He also proposed a set of patches for testing that includes a library for integrating ML models into the kernel and an example of a symbolic driver using the library. <\/p>\n
Integrating ML models into the kernel can be beneficial for altering the logic of subsystems based on the processed data, optimizing performance, and adjusting configurations according to the internal state of the systems. The application of machine learning, which can identify patterns and make predictions without manual algorithm implementation, will simplify the selection of the most effective kernel configuration considering the complexity and variability of modern workloads, as well as addressing issues like predicting failures in storage systems. <\/p>\n
P\u00ebrve\u00e7 motorit p\u00ebr ekzekutimin e modeleve, po shqyrtohet zhvillimi i mjeteve p\u00ebr mbledhjen e t\u00eb dh\u00ebnave p\u00ebr trajnimin e modeleve ML, trajnimin e drejtp\u00ebrdrejt\u00eb t\u00eb modelit dhe testimin e rezultateve. Duke qen\u00eb se p\u00ebr ekzekutimin e nj\u00eb modeli ML k\u00ebrkohen operacione me numra me presje fluturuese, dhe n\u00eb b\u00ebrtham\u00eb nuk lejohet p\u00ebrdorimi i drejtp\u00ebrdrejt\u00eb t\u00eb FPU, prototipi i propozuar p\u00ebrb\u00ebn nj\u00eb shtres\u00eb p\u00ebr qasjen nga n\u00ebn-sistemet e ndryshme t\u00eb b\u00ebrtham\u00ebs p\u00ebr n\u00eb modelet ML, t\u00eb cilat ekzekutohen n\u00eb hap\u00ebsir\u00ebn e p\u00ebrdoruesit, ngjash\u00ebm me nxjerrjen n\u00eb hap\u00ebsir\u00ebn e p\u00ebrdoruesit t\u00eb trajtuesve SPDK, DPDK dhe ublk. <\/p>\n
D\u00ebrgimi i ekzekutimit dhe trajnimet e modelit n\u00eb hap\u00ebsir\u00ebn e p\u00ebrdoruesit e thjeshton mb\u00ebshtetje dhe izolon b\u00ebrtham\u00ebn nga problemet n\u00eb kodin e ekzekutimit t\u00eb modelit. Gjat\u00eb faz\u00ebs s\u00eb trajnimit, t\u00eb dh\u00ebnat mbi gjendjen dhe parametrat e b\u00ebrtham\u00ebs mund t\u00eb merren nga p\u00ebrpunuesi nga hap\u00ebsira e p\u00ebrdoruesit, si dhe t\u00eb d\u00ebrgohen p\u00ebrmes nj\u00eb niveli q\u00eb ekzekutohet n\u00eb nivelin e b\u00ebrtham\u00ebs. P\u00ebr t\u00eb menaxhuar nd\u00ebrveprimin e komponent\u00ebve n\u00eb b\u00ebrtham\u00eb dhe hap\u00ebsir\u00ebn e p\u00ebrdoruesit, aplikohet sysfs. \u00cbsht\u00eb e mundur t\u00eb b\u00ebhet trajnimi adaptiv i modelit ML, ku n\u00ebn-sistemi i b\u00ebrtham\u00ebs merr rekomandimin nga modeli ML, aplikon ndryshimin e rekomanduar dhe e vler\u00ebson efikasitetin e rekomandimit p\u00ebr ndryshimin e gjendjes. \n Burimi: opennet.ru<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"