Ihe eserese Seldon Core na Inference
Seldon Core bụ ikpo okwu mepere emepe maka ibunye ụdị mmụta igwe na Kubernetes, nwere njirimara pụtara ìhè: eserese nrụbama.
Nchịkọta
Instead of serving one isolated model, it lets you chain models, routers, combiners, and transformers into a single directed graph that runs as one deployable service.
Ime miri emi
Ọtụtụ ezigbo mmepụta ojiji ikpe gụnyere ihe karịrị otu oku nlereanya. Ị nwere ike hazie ntinye, mee arịrịọ ka otu n'ime ụdị dị iche iche, mee mkpokọta, wee megharịa nsonaazụ ya. Seldon Core na-akọwa nke a dị ka eserese ntinye aka akọwapụtara na SeldonDeployment (ma ọ bụ, na v2 architecture, site na Seldon Core Operator na MLServer). Ewubere eserese ahụ site na ụdị akụrụngwa eji emeghari: Model na-eje ozi amụma, onye ntụgharị na-agbanwe ntinye ma ọ bụ ntinye, onye na-ahụ maka ụzọ na-ekpebi nwa ọ ga-akpọ (na-enyere ule A / B na ndị ohi nwere ọtụtụ ogwe aka), yana Combiner na-achịkọta ihe sitere na ọtụtụ ụdị maka ịgbakọta. Seldon na-akwado ọtụtụ usoro site na ihe nkesa akwadoro na ihe mkpuchi Python omenala, ọ na-ekpughekwa metrics bara ụba, ịchụ nta nkesa, na ịpụ n'ime igbe maka nleba anya na nkọwa.
Nghọta nka nka
Eserese inference bụ eserese acyclic a na-eduzi ebe ọnụ ọ bụla bụ microservice nwere ọkọlọtọ amụma amụma, yana onye na-agụ egwú Seldon (onye na-ahụ maka ọrụ / onye ọrụ) na-aga arịrịọ site na eserese ahụ wee jikọta nzaghachi. N'ihi na ndị na-anya ụgbọ ala nwere ike mejuputa mgbagha mgbagha nwere ọtụtụ ogwe aka, okporo ụzọ nwere ike gbanwee n'ụzọ dabara adaba na ụdị na-eme nke ọma dabere na akara ụgwọ ọrụ ndụ. Seldon Core v2 wepụrụ eserese ahụ site na ihe nkesa ụdị n'otu n'otu na-eji MLServer na Protocol mepere emepe, na-eme ka ọtụtụ ihe nlereanya na-eje ozi na gafere na ngwaike nkekọrịta.
Mmetụta atụmatụ
Ọnụ ego na mmefu ego
Mkpebi ihe owuwu ụlọ na-akwalite arụmọrụ yana ọnụ ahịa ọrụ ruo ọtụtụ afọ.
Mkpebi doro anya
Nkà mmụta nka na-enyere ndị otu egwuregwu aka ịhọrọ nchịkọta ziri ezi, ọ bụghị naanị nke kachasị ọhụrụ.
Quality akara
Nhọrọ injinia ka mma na-ebelata ihe omume ntụkwasị obi na mmepụta.
Ọdịnihu nke Seldon Core na eserese eserese
Seldon na-aga n'ihu na modular, data-centric MLOps nwere pipeline Core v2 na nhazi data, gbakwunyere njikọ siri ike na nchọpụta mkpagharị (Alibi Detect) na nkọwa (Alibi Kọwaa). Dị ka LLMs na sistemu na-arụ ọrụ na-aghọ eserese dị iche iche nke iweghachite, ụdị, na ngwaọrụ, maapụ abstraction nke ntinye aka na-adabere na usoro ọrụ ndị a. Na-atụ anya mkpasi ike karịa na ọtụtụ ihe nlere na-eje ozi arụmọrụ, nkwanye, na nleba anya ahaziri ahazi ka usoro AI dị mgbagwoju anya, ọtụtụ nzọụkwụ wee bụrụ nke a na-emegharị emegharị na nke a na-achịkwa na mmepụta.
Mmejuputa n'ezie n'ụwa
Onye na-agbazinye ihe na-ekenye ihe ngbanwe nke otu ọkụ na-etinye atụmatụ n'ime ọnụ ihe nlereanya, wee bụrụ Transformer na-ahazi akara ahụ, ha niile dị ka SeldonDeployment.
Otu ụlọ ọrụ mgbasa ozi na-eji ọnụ ụzọ Router na-agba ndị omempụ nwere ọtụtụ ngwa iji ziga ọtụtụ okporo ụzọ na ụdị nkwanye ọ bụla na-enweta ụgwọ ọrụ pịa dị elu.
Otu otu na-ejikọta ụdị aghụghọ atọ nwere ọnụ ọnụ Combiner nke na-anata akara ha tupu ha alaghachi otu mkpebi nye onye na-akpọ oku.
Onye na-ahụ maka inshọransị ahaziri na-etinye ndekọ ụgwọ ọrụ Seldon yana ndị na-akọwa Alibi na eserese ntinye ka enwere ike ịchọta amụma ọ bụla wee kọwaa ya maka nyocha.
Ihe ize ndụ & okporo ụzọ nche
Ịkwalite otu akara ngosi nwere ike zoo adịghị ike sistemụ sara mbara.
A na-eledakarị ihe akụrụngwa na ụgwọ ọrụ anya.
Ọdịiche nchekwa na nleba anya nwere ike itolite ka sistemu na-adịwanye mgbagwoju anya.
Map mmejuputa
Kọwaa latency, ịdịmma na ebumnuche ọnụ ahịa tupu mmejuputa ya.
Benchmark n'okpuru ibu dị adị na ọnọdụ data.
Nleba anya akụrụngwa maka mperi, ịkpafu na mmetụta onye ọrụ.
Kwadebe ụzọ nzaghachi azụghachi azụ na ihe omume tupu ịchachaa.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Seldon Core and Inference Graphs quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
TensorRT na Injin Inference
Ajụjụ a na-ajụkarị
What is Seldon Core and Inference Graphs?
Seldon Core bụ ikpo okwu mepere emepe maka ibunye ụdị mmụta igwe na Kubernetes, nwere njirimara pụtara ìhè: eserese nrụbama. Kama ijere otu ụdị dịpụrụ adịpụ ozi, ọ na-ahapụ gị ka ị yinye ụdị, ndị na-anya ụgbọ mmiri, ndị na-agwakọta ihe, na ihe ntụgharị ka ọ bụrụ otu eserese eduzi nke na-arụ ọrụ dị ka otu ọrụ nwere ike ibugharị.
Kedu ihe na-akọwapụta ọdịiche dị na Seldon Core na sava otu ụdị?
Seldon Core na-ahapụ gị ka ịdepụta ụdị, ndị na-anya ụgbọ mmiri, ndị na-agwakọta ihe, na ndị ntụgharị ka ọ bụrụ otu eserese agbakwunyere dị ka otu ọrụ.
Kedu ihe eserese Seldon na-ekpebi ọnụ ọnụ nwata ka ọ ga-ezigara arịrịọ, na-enyere ule A/B aka?
Onye rawụta na-eziga arịrịọ n'aka otu n'ime ụmụ ya ma nwee ike mejuputa ule A/B ma ọ bụ ndị ohi nwere ọtụtụ ihe.
Kedu ụdị akụrụngwa na-achịkọta nsonaazụ sitere na ọtụtụ ụdị maka ịgbakọta?
Ngwakọta ọnụ na-ejikọta nzaghachi nke ọtụtụ ụdị ụmụaka n'ime otu mmepụta, nke bụ otú e si ewuo ensembles.
Kedu ụdị nhazi eserese nke Seldon na-etolite eserese eserese?
A na-eduzi eserese nke Seldon eserese acyclic ebe ọnụ nke ọ bụla bụ microservice nwere ọkọlọtọ amụma amụma.
Na Seldon Core v2, kedu ihe nkesa na ụkpụrụ na-enyere ọtụtụ ihe nlereanya na-eje ozi n'ofe eserese ahụ?
Core v2 na-ewepụ eserese ahụ site na sava ihe nlereanya na-eji MLServer yana Protocol Mepee Inference maka ijere ọtụtụ ihe nlereanya.