Iyini i-Neural Engine kuma-Apple Silicon Macs?

Lapho i-Apple isuka kuma-Intel processors iye kwisixazululo sayo sama-chips e-Apple Silicon kumakhompyutha ayo, yathuthukisa kakhulu ukusebenza nokusetshenziswa kwamandla. Ngisho nangesikhathi sesethulo ngokwaso, wagqamisa amaphrosesa ayinhloko, ahlanganisa ndawonye akha i-chip yonke futhi angemuva kwamakhono ayo. Yiqiniso, kulokhu sisho CPU, GPU, Neural Engine nabanye. Nakuba indima ye-CPU ne-GPU ngokuvamile yaziwa, abanye abasebenzisi be-Apple abakacaci ukuthi empeleni iNeural Engine isetshenziselwe ini.

I-Cupertino giant e-Apple Silicon isekelwe kuma-chips ayo e-iPhone (A-Series), afakwe cishe ama-processor afanayo, kuhlanganise ne-Neural Engin eshiwo ngenhla. Kodwa-ke, akukho ngisho neyodwa idivaysi ecace bha ukuthi isetshenziselwa ini nokuthi kungani siyidinga nhlobo. Yize sikucacele lokhu nge-CPU ne-GPU, le ngxenye ifihlekile, kuyilapho iqinisekisa izinqubo ezibalulekile ngemuva.

Kungani kukuhle ukuba ne-Neural Engine

Kodwa ake sikhanyise mayelana nokubalulekile, noma okuhle ngempela, ukuthi ama-Macs ethu anama-Apple Silicon chips afakwe i-Neural Engine processor ekhethekile. Njengoba kungenzeka uyazi, lesi sigaba siqondene ngqo nokusebenza ngobuhlakani bokwenziwa nokufunda komshini. Kodwa lokho kukodwa akumele kuveze okuningi. Kodwa-ke, uma singakufingqa ngokujwayelekile, singasho ukuthi iphrosesa isebenzela ukusheshisa imisebenzi efanele, okwenza umsebenzi we-GPU yakudala ubonakale kalula futhi isheshisa wonke umsebenzi wethu kukhompuyutha esinikeziwe.

Lena yindlela i-Apple eyethule ngayo i-chip ye-M1

mpv-shot0102

mpv-shot0104

mpv-shot0106

mpv-shot0108

mpv-shot0109

mpv-shot0113

mpv-shot0115

mpv-shot0117

Ngokuqondile, i-Neural Engine isetshenziselwa imisebenzi ehlobene, okuthi, uma uthi nhlá, ingahlukanga nganoma iyiphi indlela kwejwayelekile. Lokhu kungaba ukuhlaziywa kwevidiyo noma ukubonwa kwezwi. Ezimweni ezinjalo, ukufunda ngomshini kuyangena, okuyinto edinga ngokuqondakalayo ekusebenzeni nasekusetshenzisweni kwamandla. Ngakho-ke akulimazi neze ukuba nomsizi osebenzayo ogxile kakhulu kulolu daba.

I-chip ye-M1 kanye nezingxenye zayo eziyinhloko

Ukusebenzisana ne-Core ML

Uhlaka lwe-Apple's Core ML nalo luhambisana neprosesa ngokwayo. Ngayo, abathuthukisi bangasebenza ngamamodeli okufunda omshini futhi badale izinhlelo zokusebenza ezithokozisayo ezizobese zisebenzisa zonke izinsiza ezitholakalayo ekusebenzeni kwazo. Kuma-iPhones wesimanje nama-Mac anama-Apple Silicon chips, i-Neural Engine izowasiza kulokhu. Ngemuva kwakho konke, lesi futhi ngesinye sezizathu (hhayi kuphela) ukuthi kungani ama-Mac emuhle futhi enamandla endaweni yokusebenza ngevidiyo. Esimeni esinjalo, abathembeli kuphela ekusebenzeni kwe-graphics processor, kodwa futhi bathola usizo oluvela ku-Neural Engine noma ezinye izinjini zemidiya zokusheshisa ividiyo ye-ProRes.

Uhlaka lwe-Core ML lokufunda ngomshini lusetshenziswa ezinhlelweni ezahlukahlukene

I-Neural Engine ekusebenzeni

Ngenhla, sesivele sidweba kancane ukuthi empeleni iNeural Engine isetshenziselwe ini. Ngokungeziwe kuzinhlelo zokusebenza ezisebenza ngokufunda komshini, izinhlelo zokuhlela amavidiyo noma ukuqashelwa kwezwi, sizokwamukela amandla ako, ngokwesibonelo, kuhlelo lokusebenza lwendabuko Izithombe. Uma usebenzisa umsebenzi Wombhalo Obukhoma ngezikhathi ezithile, lapho ungakopisha khona umbhalo obhaliwe kunoma yisiphi isithombe, i-Neural Engine ingemuva kwawo.