BentoML na Nkwakọ ngwaahịa Model
BentoML bụ usoro Python mepere emepe nke na-achịkọta ụdị mmụta igwe a zụrụ azụ ka ọ bụrụ nkeji ahaziri ahazi, nke enwere ike ibugharị akpọrọ 'Bentos'.
Nchịkọta
It bridges the gap between a model sitting in a notebook and a production service that can actually serve predictions over an API.
Ime miri emi
Mgbe onye ọkà mmụta sayensị data gụchara ọzụzụ ihe atụ, itinye ya n'ime mmepụta na-apụtakarị iji aka dee koodu ozi, ịdabere na ya, wuo onyonyo Docker, yana ịwụnye API. BentoML na-akpaghị aka nke a. Ị na-echekwa ihe nlereanya na ụlọ ahịa ihe ngosi mpaghara ya, wee kọwapụta klas ọrụ nwere njedebe API a chọrọ mma iji jikwaa ntule. Iwu 'bentoml wu' na-achịkọta ihe nlereanya, koodu Python gị, ụdị ndabere, yana nhazi oge n'ime Bento nke nwere onwe ya. Site n'ebe ahụ 'bentoml containerize' na-emepụta ihe oyiyi OCI Docker. BentoML na-akwado ihe fọrọ nke nta ka ọ bụrụ usoro ọ bụla (PyTorch, TensorFlow, scikit-learn, XGBoost, Hugging Face Transformers, ONNX) ma na-agbakwunye micro-batching na-agbanwe agbanwe, nke na-achịkọta arịrịọ mbata na-akpaghị aka iji bulie mmepụta GPU na-enweghị ịgbanwe koodu gị.
Nghọta nka nka
BentoML na-ekewapụ 'Ndị na-agba ọsọ' (mmebi ihe nlere dị arọ) na mgbagha ihe nkesa API. Ndị na-agba ọsọ nwere ike ịgba ọsọ n'onwe ha ma na-agba ọsọ na usoro ndị ọrụ nke ha, ebe ihe nkesa HTTP/gRPC dị arọ na-ejikwa arịrịọ routing na I/O. Batching ya na-emegharị emegharị na-emegharị nha batch na windo nkwụsị n'oge oge, ya mere ọ na-amịkọrọ okporo ụzọ gbawara ma na-eme ka ndị na-eme ngwa ngwa dị oke ọnụ. Ụdị Bento ahaziri ahazi na-etinye ihe ngosi, faịlụ ihe nlere anya na gburugburu ebe a na-emegharịgharị, na-eme ka ọ na-ewuli elu n'ofe igwe.
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 BentoML na Nkwakọ ngwaahịa Model
BentoML dabere n'ụdị asụsụ buru ibu yana ọrụ AI na-emepe emepe, yana OpenLLM na BentoCloud na-enye nzaghachi token nkwanye, autoscaling, na nhazi oge GPU maara. Na-atụ anya njikọta siri ike na ndị na-ebubata ihe dị ka vLLM na TensorRT-LLM, nkwado ka mma maka sistemu AI nke ọtụtụ ụdị, yana ụzọ dị nro site na ngwugwu Bento ruo na mbugharị GPU enweghị nkesa. Ka ndị otu na-esi na otu ụdị na-esi na pipeline na-arụ ọrụ, BentoML na-edobe onwe ya dị ka nkwakọ ngwaahịa na ihe na-eje ozi na-ejikọta ihe ndị ahụ ọnụ.
Mmejuputa n'ezie n'ụwa
Otu ndị na-achọpụta aghụghọ na-echekwa ụdị XGBoost na ụlọ ahịa BentoML wee wuo Bento nke na-ekpughe njedebe REST / amụma maka ọrụ ịkwụ ụgwọ ịkpọ ozugbo.
Otu ikpo okwu ML na-eji 'bentoml containerize' iji tụgharịa ụdị mmetụta ihu ịmakụ ka ọ bụrụ onyonyo Docker nke na-ebuga na ụyọkọ Kubernetes dị n'ime ha.
Mmalite na-enye ụdị Llama nke ọma nke nwere OpenLLM (wuru na BentoML), na-ebunye akara ngosi na UI nkata nwere batching mmegharị na-edobe GPU juputara.
Otu ụlọ ọrụ na-ahụ maka kọmpụta na-achịkọta ihe osise PyTorch yana pipeline na-ahazi ya ka ọ bụrụ otu Bento ka ọ bụrụ mgbanwe mgbanwe ejiri na ụgbọ mmiri na-azụ ihe nlereanya.
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 BentoML and Model Packaging 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
Ndozi nkọwa maka ụdị koodu
Ajụjụ a na-ajụkarị
What is BentoML and Model Packaging?
BentoML bụ usoro Python mepere emepe nke na-achịkọta ụdị mmụta igwe a zụrụ azụ ka ọ bụrụ nkeji ahaziri ahazi, nke enwere ike ibugharị akpọrọ 'Bentos'. Ọ na-ejikọta ọdịiche dị n'etiti ihe atụ nọ ọdụ na akwụkwọ ndetu na ọrụ mmepụta nke nwere ike ibu amụma n'ezie na API.
Kedu otu ahaziri ahazi, nke enwere ike ibugharị nke BentoML na-emepụta?
BentoML na-achịkọta ihe nlereanya, koodu ya, ndabere ya, na nhazi oge ojiri gaa n'ime ụdị nke nwere onwe ya nke akpọrọ Bento.
Kedu ihe BentoML's micro-batching na-eme ka ọ dịkwuo mma?
Arịrịọ na-abata ndị otu batching na-eme mgbanwe n'oge a na-agba ọsọ iji mee ka GPU na-arụ ọrụ ma na-ebuli mmepụta na-enweghị mgbanwe koodu.
N'ime ihe owuwu nke BentoML, kedu ihe na-eji usoro ogbugbu dị arọ dị iche na sava API?
Ndị na-agba ọsọ na-echikota ntinye ihe nlere anya ma nwee ike itule n'ime usoro ndị ọrụ nke ha, ejikọtaghị ya na sava API dị fechaa.
Kedu iwu na-atụgharị Bento wuru ka ọ bụrụ onyonyo OCI Docker?
'bentoml containerize' na-ewepụta onyonyo ọkọlọtọ OCI/Docker site na Bento maka ibuga.
Kedu n'ime ihe ndị a bụ isi ihe kpatara BentoML ji tinye ụdị ndabere na Bento?
Ntunye na itinye gburugburu ebe obibi na-eme ka ọrụ ngwugwu a na-emegharịgharị ma na-agbanwe agbanwe ebe ọ bụla ọ na-aga.