BentoML da Model Packaging
BentoML shine tsarin Python mai buɗewa wanda ke haɗa samfuran injunan koyan injuna zuwa daidaitattun raka'a, da za a iya turawa da ake kira 'Bentos'.
Dubawa
It bridges the gap between a model sitting in a notebook and a production service that can actually serve predictions over an API.
Zurfafa nutsewa
Lokacin da masanin kimiyyar bayanai ya gama horar da ƙira, shigar dashi cikin samarwa yawanci yana nufin rubuta lambar sabis da hannu, haɗa abubuwan dogaro, gina hoton Docker, da haɗa API. BentoML yana sarrafa wannan. Kuna ajiye samfuri zuwa kantin sayar da samfurin gida, sannan ku ayyana ajin Sabis tare da ƙarshen API wanda aka ƙawata don sarrafa tunani. Umurnin 'bentoml gini' ya ƙunshi samfurin, lambar Python ku, nau'ikan dogaro, da daidaitawar lokacin aiki zuwa cikin abin da ya ƙunshi kansa, sigar Bento. Daga can 'bentoml containerize' yana samar da hoton OCI Docker. BentoML yana goyan bayan kusan kowane tsarin (PyTorch, TensorFlow, scikit-learn, XGBoost, Hugging Face Transformers, ONNX) kuma yana ƙara ƙaramar ƙaramar batching, waɗanda ke haɗa buƙatun masu shigowa ta atomatik don haɓaka kayan aikin GPU ba tare da canza lambar ku ba.
Fahimtar Fasaha
BentoML yana raba 'Masu gudu' (kisa-ƙirar ƙirar ƙira) daga dabarar uwar garken API. Masu gudu za su iya yin sikelin kansu da kansu kuma suna gudanar da ayyukan ma'aikatansu, yayin da uwar garken HTTP/gRPC mai nauyi ke sarrafa buƙatun routing da I/O. Batching ɗin sa na daidaitawa yana jujjuya girman juzu'i da taga latency a lokacin aiki, don haka yana ɗaukar fashewar zirga-zirgar ababen hawa kuma yana sa masu saurin gudu masu tsada. Madaidaicin tsarin Bento yana ƙunshe da bayyanuwa, fayilolin ƙira, da mahalli da za'a iya sakewa, yana gina ƙayyadaddun ƙayyadaddun injuna.
Dabarun Tasiri
Kudin da kasafin kuɗi
Hukunce-hukuncen gine-gine suna haifar da aiki da tsadar aiki na shekaru.
Shawarwari masu haske
Ilimin fasaha yana taimaka wa ƙungiyoyi su zaɓi tari mai kyau, ba kawai sabon abu ba.
Kula da inganci
Zaɓuɓɓukan injiniya mafi kyau suna rage abin dogaro a cikin samarwa.
Makomar BentoML da Model Packaging
BentoML ya jingina da ƙarfi cikin babban ƙirar harshe da kuma samar da sabis na AI, tare da OpenLLM da BentoCloud suna ba da amsoshi masu yawo, autoscaling, da kuma tsarin GPU-sane. Yi tsammanin haɗin kai mai ƙarfi tare da inganta haɓakawa kamar vLLM da TensorRT-LLM, mafi kyawun tallafi don tsarin ƙirar AI mai yawa, da kuma santsin hanyoyi daga fakitin Bento zuwa jigilar GPU mara sabar. Yayin da ƙungiyoyi ke motsawa daga ƙira ɗaya zuwa bututun mai, BentoML yana sanya kanta azaman marufi da layin hidima wanda ke haɗa waɗannan abubuwan tare.
Aiwatar da Gaskiyar Duniya
Ƙungiyar gano zamba tana adana samfurin XGBoost zuwa kantin BentoML kuma ta gina Bento wanda ke fallasa ƙarshen ƙarshen REST don sabis na biyan kuɗi don kira a ainihin lokacin.
Ƙungiyar dandalin ML tana amfani da 'bentoml containerize' don juya samfurin Hugging Fuska zuwa hoton Docker wanda ke tura gungu na Kubernetes na ciki.
Farawa yana ba da ingantaccen tsarin Llama mai kyau tare da OpenLLM (wanda aka gina akan BentoML), alamun yawo zuwa UI taɗi tare da daidaita daidaitawar GPU.
Wani kamfani mai hangen nesa na kwamfuta yana haɗa nau'in nau'in hoto na PyTorch tare da sarrafa bututunsa zuwa Bento ɗaya don haka ainihin sauye-sauyen da aka yi amfani da su wajen horar da jirgin ruwa tare da samfurin.
Hatsari & Tsare-tsare
Haɓaka ma'auni ɗaya na iya ɓoye manyan raunin tsarin.
Sau da yawa ana raina kayan more rayuwa da kuma kuɗin kulawa.
Tsaro da gibin lura na iya girma yayin da tsarin ke ƙara haɓaka.
Taswirar Hanya
Ƙayyade latency, inganci, da maƙasudin farashi kafin aiwatarwa.
Alamar ma'auni a ƙarƙashin ainihin kaya da yanayin bayanai.
Kula da kayan aiki don kurakurai, ɗigo, da tasirin mai amfani.
Shirya bijirowa da hanyoyin mayar da martani kafin sikeli.
Ci gaba da Bincike
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
Jagora na gaba
Gyaran Hasashen don Samfuran Code
Tambayoyin da ake yawan yi
What is BentoML and Model Packaging?
BentoML shine tsarin Python mai buɗewa wanda ke haɗa samfuran injunan koyan injuna zuwa daidaitattun raka'a, da za a iya turawa da ake kira 'Bentos'. Yana cike gibin da ke tsakanin samfurin da ke zaune a cikin littafin rubutu da sabis na samarwa wanda zai iya ba da tsinkaya a zahiri akan API.
Menene daidaitattun, rukunin da ake iya turawa wanda BentoML ke samarwa?
BentoML yana haɗa samfuri, lambar sa, abubuwan dogaro, da daidaitawar lokacin aiki cikin siga, naúrar da ke ƙunshe da kai da ake kira Bento.
Menene madaidaicin ƙaramar batching na BentoML ya inganta da farko?
Ƙungiyoyin batches masu dacewa da buƙatun shigowa a lokacin aiki don ci gaba da shagaltar da GPUs da haɓaka kayan aiki ba tare da canje-canje na lamba ba.
A cikin gine-ginen BentoML, menene ke aiwatar da kisa mai nauyi daban da sabar API?
Masu gudu suna ƙaddamar da ƙididdiga na ƙira kuma suna iya ƙima a cikin matakan ma'aikatan su, waɗanda aka raba su daga sabar API mai nauyi.
Wane umurni ne ke juya ginannen Bento zuwa hoton OCI Docker?
'bentoml containerize' yana samar da daidaitaccen hoton OCI/Docker daga Bento don turawa.
Wanne daga cikin waɗannan shine babban dalilin BentoML ya haɗa nau'ikan dogaro a cikin Bento?
Maƙala da haɗa mahalli yana sa sabis ɗin da aka ƙulla shi ya zama mai yiwuwa kuma ya daidaita duk inda yake gudana.