Ithuluzi Lokuqonda: Umthombo Ovulekile Wokufunda Okujulile SW

Ithuluzi Lokuqonda: Umthombo Ovulekile Wokufunda Okujulile SW

Ithuluzi Lokuqonda: I-Microsoft Open Source Deep Learning SW

I-Microsoft Cognitive Toolkit (phambilini ebibizwa nge-CNTK) iyi- amathuluzi wokufunda ajulile (Machine Learning) de «Código Abierto» ngamandla amakhulu. Futhi kumahhala, kulula ukuyisebenzisa, futhi ine- ikhwalithi yebanga lokuhweba lokho kuvumela ukudala ubuchule bokufunda obujulile obukwazi ukufunda ezingeni eliseduze nelobuchopho bomuntu.

IMicrosoft, umakhi wayo, iqinisekisa ukuthi leli thuluzi lomthombo ovulekile linikeza ukukala okungayekethisi, isivinini sekhwalithi enhle nokunemba, nokuhambisana nezilimi ezisetshenziswayo ezisetshenziswa kakhulu namuhla.

Ithuluzi Lokuqonda: Isingeniso

Futhi lokho, ngenxa yalokho, kunjalo uyakwazi ukuqeqesha nokuhlola ubuchule bokufunda obujulile ngokushesha kunamanye amathuluzi afanayo. Ngenkathi, nayo, ikwazi ukukala kahle ezindaweni ezahlukahlukene (ama-CPU, ama-GPU, kanye neQoqo Lezinsimbi) ngaphandle kokulahlekelwa ukunemba.

Kuyinto yakhiwe ngobuchule obuphezulu nabafundi bokukhiqiza ukusebenza ngokuthembekile ngamasethi wedatha amakhulu. Kusetshenziswa okufanayo emikhiqizweni efana ne- I-Skype, iCortana, iBing ne-Xbox. Futhi okwamanje iyasetshenziswa embonini enkulu yezobuchwepheshe ukuthuthukisa «Inteligencia Artificial (IA)» ibanga lokuhweba. Ngaphezu kwakho konke, kwe- ukusekelwa kwe-C ++ nePython, ekuvumela ukuthi wenze ngezifiso cishe noma yikuphi okukhona noma okwakho ukuqeqeshwa kwe-algorithm.

Ikhithi Yamathuluzi Okuqonda: Okuqukethwe

Ikhithi yamathuluzi okuqonda: Kuyini futhi yiziphi izici zayo?

  • Ithuluzi lomthombo ovulekile lokufunda okujulile.
  • Kungumkhiqizo owenziwe yiMicrosoft onenguqulo yawo yakamuva engu-2.7.
  • Chaza amanethiwekhi we-neural njengochungechunge lwezinyathelo zokubala ngokusebenzisa igrafu eqondisiwe.
  • Ivumela abasebenzisi ukuthi benze kalula futhi bahlanganise izinhlobo ezithandwa kakhulu amamodeli, njenge: Deep Neural Networks, Convolutional and Recurrent.
  • Sebenzisa ukufunda okwehlayo kwe-stochastic gradient ngokwahlukanisa okuzenzakalelayo nokuqhathanisa kuma-GPU amaningi kanye namaseva.
  • Ingafakwa njengomtapo wezincwadi ezinhlelweni zePython, C # noma ze-C ++, noma isetshenziswe njengethuluzi lokufunda elizimele elijulile ngolimi lwayo oluchaza imodeli (BrainScript). Noma sebenzisa ukusebenza kwayo kokuhlola kusuka ezinhlelweni ezenziwe ngeJava.
  • Iyahambisana ne-Linux 64-bit noma i-Windows 64-bit Operating Systems ngamaphakeji kanambambili asungulwe ngaphambilini, noma ngokuhlanganiswa kusetshenziswa amafayela womthombo wawo abanjwe ku-GitHub.

Nota: Ngeminye imininingwane ngeCognitive Toolkit ungavakashela i- Iwebhusayithi esemthethweni kwaMicrosoft noma kuwebhusayithi yayo ku GitHub.

I-ONNX

El «Intercambio de Redes Neuronales Abiertas (Open Neural Network Exchange - ONNX)» i-ecosystem evulekile evumela abathuthukisi be «IA» khetha amathuluzi afanele njengoba iphrojekthi yakho iguquka.

Y Ithuluzi Lokuqonda ingelinye lamathuluzi okufunda ajulile okuqala wokuxhasa i- ifomathi «ONNX». Fometha ukuthi yini i- yenziwe ngokubambisana yiMicrosoft futhi kusekelwa abanye abaningi. Ukuze, «ONNX» ivumela abathuthukisi ukuhambisa amamodeli phakathi kwezinhlaka ezifana I-Cognitive Toolkit, Caffe2, MXNet, ne-PyTorch.

Ekugcineni, «ONNX» inikeza ifomethi yomthombo ovulekile ye- Amamodeli we «IA» bese uchaza a imodeli yokuqhafaza enwebekayo yokuqhafazafuthi kufaka phakathi izincazelo zo-opharetha abakhelwe ngaphakathi nezinhlobo zedatha ezijwayelekile. Futhi igxila kumakhono adingekayo izinqubo zokucabanga (ukushaya amagoli).

Nota: Ngeminye imininingwane nge «ONNX» Ungavakashela iwebhusayithi esemthethweni efanayo kusixhumanisi esilandelayo Vula iNeural Network Exchange.

Ikhithi Yamathuluzi Okuqonda: Isiphetho

Isiphetho

Sithemba ukuthi ukhona "okuncane kodwa okuwusizo" ngalokhu enye isoftware ethokozisayo «Código Abierto» i-del I-Microsoft Open Application Ecosystem izingcingo «Cognitive Toolkit», yiba nentshisekelo nokusetshenziswa okuhle, kwayo yonke «Comunidad de Software Libre y Código Abierto» kanye negalelo elikhulu ekusabalalisweni kwesimangaliso semvelo, esikhulu futhi esikhulayo sokusebenzisa kanye nokwenzelwe «GNU/Linux».

Futhi ukuthola eminye imininingwane, unganqikazi njalo ukuvakashela noma yikuphi Umtapo wolwazi oku-inthanethi njengoba I-OpenLibra y I-JedIT ukufunda izincwadi (ama-PDF) ngalesi sihloko noma ezinye izindawo zolwazi. Okwamanje, uma ukuthandile lokhu «publicación», ungayeki ukuyaba nabanye, ku- Amawebhusayithi ayizintandokazi, iziteshi, amaqembu, noma imiphakathi zokuxhumana nabantu, okungcono mahhala futhi evulekile njenge I-mastodon, noma ukuphepha nokuyimfihlo kwangasese yocingo.

Noma mane uvakashele ikhasi lethu lasekhaya ku- KusukaLinux noma ujoyine iSiteshi esisemthethweni Yocingo kusuka ku-DesdeLinux ukufunda nokuvotela lokhu noma ezinye izincwadi ezithakazelisayo ku- «Software Libre», «Código Abierto», «GNU/Linux» nezinye izihloko ezihlobene ne- «Informática y la Computación»kanye «Actualidad tecnológica».


Okuqukethwe yi-athikili kunamathela ezimisweni zethu ze izimiso zokuhlelela. Ukubika iphutha chofoza lapha.

Yiba ngowokuqala ukuphawula

Shiya umbono wakho

Ikheli lakho le ngeke ishicilelwe. Ezidingekayo ibhalwe nge *

*

*

  1. Ubhekele imininingwane: Miguel Ángel Gatón
  2. Inhloso yedatha: Lawula Ugaxekile, ukuphathwa kwamazwana.
  3. Ukusemthethweni: Imvume yakho
  4. Ukuxhumana kwemininingwane: Imininingwane ngeke idluliselwe kubantu besithathu ngaphandle kwesibopho esisemthethweni.
  5. Isitoreji sedatha: Idatabase ebanjwe yi-Occentus Networks (EU)
  6. Amalungelo: Nganoma yisiphi isikhathi ungakhawulela, uthole futhi ususe imininingwane yakho.