Kubeflow: Masini Aʻoaʻoga Meafaigaluega mo Kubernetes

Kubeflow: Masini Aʻoaʻoga Meafaigaluega mo Kubernetes

Kubeflow: Masini Aʻoaʻoga Meafaigaluega mo Kubernetes

O le tatou tulaga i le aso o le a feagai ma le fanua o Aʻoaʻo otometi (Masini Aʻoaʻo / ML). Faʻamatalaga faʻapitoa e uiga i se tatala tatala tusi apalai valaʻauina "Kubeflow", lea e aoga, Kubernetes. Lea, pei o le toʻatele o outou ua uma ona iloa, o se matala mafuaʻaga polokalama mo otometi le faʻatinoina, scaling ma le tagofiaina o containerized talosaga.

"Kubeflow" e ui lava o loʻo avanoa nei i lalo ole mautu lomiga 1.2, e pei ona aliali mai i lana ofisa aloaia 'upega tafaʻilagi ma GitHub, i lana aloaia Blog, ua uma ona faʻaalia luga o le isi lomiga 1.3. O le mafuaaga lena i aso nei, o le a tatou suʻesuʻe loloto i lenei tusi talosaga.

Cognitive Toolkit: Open Source Deep Aʻoaʻo SW

Cognitive Toolkit: Open Source Deep Aʻoaʻo SW

Ma e pei ona masani ai, mo i latou e naunau i taimi uma e suʻesuʻe i se mataupu faitau, o le a matou tuʻuina ia fesoʻotaʻiga i fesoʻotaʻiga muamua pou mo oe e suʻesuʻe pe a maeʻa lenei pou.

"O le Microsoft's Cognitive Toolkit (na taʻua muamua o le CNTK) o se meafaigaluega aʻoaʻo loloto (Machine Learning) de «Código Abierto» ma le tele gafatia. E maua fua, faigofie e faʻaaoga, ma tulaga lelei faʻapisinisi e faʻatagaina ai oe ona fausia ni loloto algorithms e mafai ona aʻoaʻoina i se tulaga latalata i le faiʻai o le tagata." Cognitive Toolkit: Open Source Deep Aʻoaʻo SW

Tusitusiga fesoʻotaʻi:
Cognitive Toolkit: Open Source Deep Aʻoaʻo SW

Tusitusiga fesoʻotaʻi:
.NET ma ML.NET: Microsoft Open Source Platforms
Tusitusiga fesoʻotaʻi:
TensorFlow ma Pytorch: Open Source AI Faʻavae

Kubeflow: O se Polokalama Tatala Masini Poloketi

Kubeflow: O se Polokalama Tatala Masini Poloketi

O le a le Kubeflow?

E tusa ai ma lau faʻapitonuʻu aloaia, o lenei tatala poloketi ua faʻamatalaina faʻapea:

"O se poloketi tuutoina i le faia masini aʻoaʻoina (ML) workflow deployments luga Kubernetes faigofie, feaveaʻi, ma scalable. O lona faʻamoemoe e le o le toe fausiaina o isi tautua, ae ia saunia se auala faigofie e faʻatutuina ai le sili sili tatala faʻavae punaoa mo ML i isi atinaʻe eseese. Ma soʻo se mea e tamoʻe ai Kubernetes, e mafai ona tamoʻe Kubeflow."

A o, i luga o lau 'upega tafaʻilagi i GitHub, faʻaopopo puʻupuʻu le mea lea:

"Kubeflow o le tagatanuu tulaga i le ao mo masini aʻoaʻoga gaioiga: paipa, toleniga ma le faʻaogaina."

Mai lenei, e mafai ona faigofie ona toʻesea lena, o le autu faʻamoemoe o "Kubeflow" es:

"Faʻafaigofie suʻesuʻega (ML) faʻataʻitaʻiga scaling ma le faʻaaogaina i le faigofie e mafai ai, e ala i le faʻatagaina Kubernetes faia le mea na te faia: Faigofie, toe fai, ma feaveaʻi tosoina i soʻo se eseese atinae eseese, microservices faʻaogaina ma le pulega le soʻosoʻatasi faʻatasi ma fua i luga o manaoga."

Uiga o uiga?

Faatasi ai ma uiga iloga o "Kubeflow" E mafai ona matou taʻua mea nei:

  • Aofia ai tautua e fausia ai ma faʻafoeina fesoʻotaʻiga Jupiter api. Faʻatagaina e faʻapitoa le tuʻuina atu o tutusa ma isi komipiuta punaoa e faʻafetaui i latou i manaʻoga o faʻamatalaga saienisi. Ma faʻapea, faʻafaigofie ona faʻataʻitaʻi ma alualu i luma i le lotoifale, ona tuʻuina atu lea i le ao pe a tatau ai.
  • Tuʻuina mai se tu masani TensorFlow aʻoaʻoina tagata faigaluega. Lea e mafai ona faʻaaogaina e toleni ai se faʻataʻitaʻiga ML. A faʻapitoa lava, e mafai e le Kubeflow job operator ona faʻatautaia galuega tufatufa TensorFlow. Faʻatagaina le paoa e faʻatulaga ai le faʻataʻitaʻiga pule e faʻaaoga CPU poʻo GPU, ona faʻafetaui ai lea ile tele o ituaiga fuifui.
  • Lagolagoina le TensorFlow Serving container mo le auina atu i fafo faʻataʻitaʻi o le TensorFlow faʻataʻitaʻi i Kubernetes. E le gata i lea, o Kubeflow o loʻo tuʻufaʻatasia ma Seldon Core, o se avanoa matala mo le faʻaaogaina o masini e faʻataʻitaʻia ai faʻataʻitaʻiga i luga o Kubernetes, ma le NVIDIA Triton Inferensi Server e faʻateleina ai le faʻaaogaina o le GPU peʻa faʻataʻitaʻia ML / DL faʻataʻitaʻiga.
  • Faʻaaofia tekinolosi Kubeflow Pipelines. Lea o se tali maopoopo mo le faʻasalalauina ma le faʻataʻitaʻia o galuega mulimuli-i-faaiuga ML galuega. Faʻatagaina mo vave ma faʻatuatuaina faʻataʻitaʻiga, mo le faʻatulagaina ma le faʻatusatusaina o tamoʻe, ma le iloiloina auiliili lipoti i luga o tamoʻe.
  • Ofoina se tele-faʻavae faʻavae. Talu ai, i le faʻaopopoina o le galue lelei ma TensorFlow, o le a le pine ae maua le lagolago mo PyTorch, Apache MXNet, MPI, XGBoost, Chainer, ma sili atu.

Nisi faʻamatalaga lata mai luga "Kubeflow" mafai ona maua saʻo i lau Blog aloaia.

O le a le Kubernetes?

Tuuina atu le, "Kubeflow" galue pea "Kubernetes", e aoga le faʻamatalaina e tusa ma lau oe faʻapitonuʻu aloaia o le mulimuli ane o mea nei:

"Kubernetes (K8s) o se tatalaina punaoa tulaga mo otometi le faʻatinoina, scaling, ma le puleaina o containerized talosaga."

Ma i le tulaga o, manaʻo ia loloto atili i "Kubernetes" Oe mafai ona suʻesuʻeina a matou lomiga talu ai ma lata mai fesoʻotaʻi i lalo:

Tusitusiga fesoʻotaʻi:
Kubernetes 1.19 taunuʻu mai ma le tasi le tausaga o le lagolago, TLS 1.3, faʻaleleia atili ma sili atu
Tusitusiga fesoʻotaʻi:
Docker vs Kubernetes: mea lelei ma mea le lelei

Ata lautele mo faʻaiuga faʻaiuga

Faaiuga

Matou te talosia ia lenei "aoga laiti pou" e uiga i «Kubeflow», o se manaia ma faʻaonapo nei tatala matata poloketi i le matata o loloto aʻoaʻoga, faia e faʻateleina ai le aapa atu o le tatala punaoa tulaga «Kubernetes »; e matua fiafia ma aoga, mo le atoa «Comunidad de Software Libre y Código Abierto» ma le sao tele i le faʻasalalauina o le matagofie, tele ma olaolaolaolaolaolaolaolaolaolaolaolaaga o faʻaogaina o «GNU/Linux».

Mo le taimi nei, pe afai e te fiafia i lenei publicación, Aua le tuʻu fefaasoaaʻi ma isi, i luga o lau 'upega tafaʻilagi e te fiafia iai, auala, kulupu poʻo faʻalapotopotoga o fesoʻotaʻiga i luga o le upega tafaʻilagi poʻo feau feaveaʻi, e sili ona maua fua, tatala ma / pe sili atu le saogalemu e pei o uālesiFaailogaMastodon poʻo seisi o Faigofie, sili atu.

Ma manatua e asiasi i le matou itulau itulau i «MaiLinux» e suʻesuʻe atili talafou, faʻatasi ai ma le auai i la matou auala faʻapitoa o Telegram mai DesdeLinuxA o, mo nisi faʻamatalaga, e mafai ona e asia soʻo Faletusi i luga ole laiga como OpenLibra y jedit, ia maua ma faitau tusi numera (PDF) i lenei mataupu poʻo isi.


O mataupu o le tusitusiga e tausisi ia tatou mataupu silisili o amio lelei faʻatonu. E lipotia se mea sese kiliki iinei.

Ia avea muamua ma faamatalaga

Tuʻu lau faamatalaga

o le a le lomia lou tuatusi imeli.

*

*

  1. E tali atu mo faʻamatalaga: Miguel Ángel Gatón
  2. Faamoemoega o faʻamatalaga: Pulea le SPAM, faʻamatalaga pulega.
  3. Tulaga faʻatulafonoina: Lau maliega
  4. Fesoʻotaʻiga o faʻamatalaga: O faʻamatalaga o le a le fesoʻotaʻi atu i isi vaega vagana i tulafono faʻatulafonoina.
  5. Teuina o faʻamatalaga: Faʻamaumauga tuʻufaʻatasia e Occentus Networks (EU)
  6. Aia Tatau: I soo se taimi e mafai ai ona e faʻatapulaʻaina, toe maua ma aveʻese au faʻamatalaga.

bool(moni)